IMPACT OF MULTIPLE NORMATIVE SYSTEMS ON

THE ORGANIZATIONAL PERFORMANCE OF

INTERNATIONAL JOINT VENTURES

A DISSERTATION

SUBMITTED TO THE DEPARTMENT OF CIVIL AND ENVIRONMENTAL

ENGINEERING

AND THE COMMITTEE ON GRADUATE STUDIES

OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Tamaki Horii

September 2005

© Copyright by Tamaki Horii 2005

All Rights Reserved

ii

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Professor Raymond E. Levitt (Principal Adviser)

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Professor W. Richard Scott

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Associate Professor Yan Jin

Approved for the University Committee on Graduate Studies:

iii

ABSTRACT

Research on international joint-venture (IJV) teams reveals high failure rates for projects,

due to difficulties in managing mixed-cultural teams. The goal of this dissertation

research is to understand, analyze, and model how cultural differences in IJV projects

affect team performance, using case studies and computational experimentation. This

dissertation focuses on “culturally-driven normative systems” as a key element of cultural

differences. Culturally-driven normative systems refer to conceptions of preferred or

desirable standards, not only for individual behaviors in decision-making and

communication, but also for organizational practices.

The first phase of this research attempted to understand and characterize culturally-driven

normative systems of Japanese and American firms in IJV projects through case studies.

My ethnographic studies revealed distinctive individual behavior patterns and

organizational styles for different nations. Specifically, Japanese team members show

group-based decision making and communication behaviors, while American team

members have individual-based decision-making and communication behaviors.

Additionally, Japanese project teams tend to have multiple levels of hierarchy and to be

more centralized, while American firms usually adopt a flat organization hierarchy with

more decentralized authority.

In the second phase of the research, I analyzed the impact of culturally-driven normative

systems on project team performance using the Virtual Design Team (VDT)

computational simulation model of project organizations. VDT was developed and

calibrated based on information processing views (Thompson, 1967; Galbraith, 1973;

1974) to predict project organization performance such as project duration, cost, and

quality risk. However, VDT was not intended to model multiple cultures. The goal of

this phase was to explore to what extent the VDT model with the information processing

view of organizations could capture cultural impact on the performance of project

organizations. To use VDT to capture cultural differences in organization behavior, I

iv

developed a culture-organization model that maps Hofstede’s national culture attributes

into individual behavior patterns and organization styles. The design of the VDT based

simulation experiments uses the independent parameters of task complexity at four levels

and team experience at three levels, yielding twelve organizational contexts. I simulated

the four possible combinations of US vs. Japanese team individual behavior patterns and

US vs. Japanese organization styles in each context to predict project duration, work

volume, cost, and quality outcomes. The simulation results indicated that: 1) both

Japanese and American teams show better performance across all contexts when each

employs its familiar organization style; 2) the Japanese organization style generally

performs better under high task complexity, while the American organization style

generally performs better under medium and low task complexity; and 3) the Japanese

organization style generally leads to lower project quality risk than the American

organization style. In addition, culturally-driven individual behavior patterns have less

impact on project outcomes than culturally-driven organization styles. This experiment

showed that VDT’s information processing model of project organizations can be applied

to capture cultural impact on project performance in single-cultural teams. However,

VDT in its current form cannot be used to model the interactions between different

cultures found in mixed-cultural project teams.

Finally, to address the limitations of VDT in modeling culturally-driven organization

styles and decision-making behaviors in mixed-cultural teams, in the third phase of the

research I developed a prototype computational model — “InterCultural-Virtual Design

Team” (IC-VDT) — to analyze the impacts of cultural interactions on the performance of

mixed-culture teams. IC-VDT can represent and reason about multiple individual

behavior patterns in a project. IC-VDT also incorporates exceptions caused by differing

organization practices — called institutional exceptions. IC-VDT was developed based

on my field observations of US-Japanese joint venture projects in the US, and on

descriptions of Japanese vs. US organizational practices from the literature. Using IC-

VDT, I examined the impact of mixed-cultural teams on project performance with the

same organizational contexts. Simulated results shows that: (1) the American

organization style reduces the negative impacts of having mixed-cultural teams compared

v

to the Japanese organization style, while the Japanese organization style shows better

performance in project quality risks; (2) Mixed-cultural teams can be less efficient (i.e.,

70% worse) or slightly more efficient (i.e., 30% improved) than single-cultural teams;

and (3) high matrix strength improves both project duration and project work volume for

mixed-cultural teams, but does not affect project or functional quality risks.

The simulation results are qualitatively consistent with organizational contingency theory,

cultural contingency theory, and limited observations of US-Japanese IJV project teams,

thus providing initial validation for the reasoning of IC-VDT. These results extend the

possibility of using simulation modeling to capture distinguishing cross-cultural

phenomena that emerge in global projects.

Key Words: Cultural differences; practices; values; Japanese; American; Virtual Design

Team (VDT); project performance; organization design; institutions; cultural interactions;

computational modeling; organizational simulation, computational social science, mono-

cultural; cross-cultural; intercultural.

vi

ACKNOWLEDGEMENTS

As with any project of this magnitude, I owe a debt of gratitude to many people for their

help along the way. In particular, without the friendship and intellectual support of my

advisors, Professor Raymond E. Levitt, Professor Yan Jin, and Professor W. Richard

Scott, I might never have reached the conclusion of my PhD work. Their enthusiasm,

patience, and keen insights have made Stanford graduate school an interesting and

enjoyable experience. Additionally, many thanks go to Professor Rich Burton at the

Fuqua School of Business at Duke University, who, despite the long distance from

Stanford University, actively gave me many insights and much support for my research.

I would like to express my gratitude to those members of the Stanford faculty who

have helped me with my research. Professor Martin Fischer, Prof. Stephen R. Barley,

Prof. John Kunz, Prof. Russell Clough, Prof. Melody Spladlin, Prof. Befu Harumi, Prof.

Aoki Masahiko, and Dr. Phil Herbert have all provided new and useful insights.

I also would like to express my appreciation to the interviewees who provided

precious data as well as comments. I am indebted for their numerous contributions. In

particular, Mick Miyake, Mozan Totani, Dr. Bryan Moser, Dr. Makoto Kataoka, Sato

Takuzo, Imai Tomoya participated actively in the Japanese-American business study

group, and gave me numerous insights.

Many thanks also go to CRGP (Collaboratory for Research on Global Projects)

group members: Dr. Julie Kim, Joyce Kiefer, Ashwin Mahalingam, Ryan Orr, John

Taylor, Rahinah Binti Ibrahim, Johanna Nummelin, Sampo Tukiainen, and Dr. Tapio

Koivu. Without these friends, my life here would have been dull indeed.

My appreciation extends to generous financial supports from the National Science

Foundation under Grant No. IIS-9907403, the Clarkson H. Oglesby Memorial Fellowship

Fund, industrial affiliates of Collaboratory for Research on Global Projects (CRGP) <

http://crgp.stanford.edu >. In addition, many thanks go to Obayashi Corporation and its

staff for providing me the opportunity to attend Stanford as a graduate student.

I am greatly indebted to Dr. Carol Cain, who very kindly undertook the difficult

task of correcting my English. I was very excited that she was able to make my

vii

manuscript so much more readable, without altering even a minor point in the flow of my

discussion.

Finally, my deepest gratitude and appreciation is extended to my family, Yoshiko

and Tatsuki, for their faith and patience. I cannot count how many times I was cheered

up by their smiles and kind words. Without their support, I could not have reached this

goal at all.

viii

Table of Contents 1. Introduction 1

1.1 Purpose and Motivation 1

2. Points of Departure 7 2.1 Culture and Institutional Theory 7

2.1.1 What is “Culture”? 2.1.2 Cultural Values 2.1.3 Cultural Practices 2.1.4 Institutional Theory

2.2 Cross-Cultural and Intercultural Research on Japanese and American Corporations 19 2.2.1 Cross-Cultural Research 2.2.2 Intercultural Research

2.3 Organization Theory 24

3. Objectives and Approach 28 3.1 Research Objectives and Research Questions 28 3.2 Research Approach 29 3.3 Research Steps 31 3.4 Research and Validation Process 33

4. Case Study 37 4.1 Case Study 37 4.2 Methodology 40 4.3 Observations 42

4.3.1 Value Differences 4.3.2 Practice Differences 4.3.3 Others

4.4 Conclusion and Discussion 59

5. Intellective Experiments for Single-Cultural Teams 63 5.1 Simulation Models as a Methodology 64 5.2 Modeling 65

5.2.1 Organization Structure 5.2.2 Individual Behavior Pattern 5.2.3 Task Complexity 5.2.4 Team Experience

5.3 Experimentation 85 5.3.1 Settings for Intellective Experiment 5.3.2 Simulated Results 5.3.3 Components of Hidden Work

ix

5.4 Discussion and Conclusion 101

5.4.1 Implications 5.4.2 Validity and Limitations 5.4.3 Conclusion

6. Prototype Model, InterCultural-Virtual Design Team 108

6.1 Purposes and Agenda of the IC-VDT Model 108 6.2 Framework for the IC-VDT Model 109

6.2.1 Value Differences 6.2.2 Practice Differences

6.3 Contingency Fit 120 6.4 Summary of the IC-VDT Model 122

6.4.1 Overview 6.4.2 Implementation of IC-VDT

7. Intellective Experiments for Mixed Cultural Teams 126

7.1 Hypotheses for Managing Mixed-Cultural Teams 126 7.2 Parameters Used for Intellective Experiments 130 7.3 Experimentation 134

7.3.1 Simulated Results 7.3.2 Statistical Analyses

7.4 Discussion and Conclusion 154 7.4.1 Implications 7.4.2 Validation 7.4.3 Limitations 7.4.4 Conclusion

8. Conclusion and Contributions 167

8.1 Conclusion 167 8.2 Contributions 171

8.2.1 Contributions to Organization Science 8.2.2 Contributions to Cultural and Institutional Theory 8.2.3 Contributions to Cross-Cultural and Intercultural Research on

Japanese and American Corporations 8.2.4 Contributions to Practice

8.3 Future Research 176 Appendix A: American Behavior Pattern 178 Appendix B: Japanese Behavior Pattern 180 References 182

x

List of Tables Chapter 2 2.1 Raw Score Along Each Dimension for Japan and the USA 12 2.2 Institutional Elements and Carriers 16 Chapter 4 4.1 Description of Case Studies 39 4.2 Summary of Interviews – Case Study 1: SC Project 51 4.3 Summary of Interviews – Case Study 2: C Bridge Project 54 4.4 Summary of Interviews – Case Study 3: G Bridge Project 57 4.5 Summary of Interviews – Case Study 4: SF Tunnel Project 58 4.6 Summary of Findings 61

Chapter 5 5.1 Leadership Styles as Organization Structure 68 5.2 National Cultural Index – Behavior Matrix 69 5.3 Adjustment Factor for Decision Making Policy 72 5.4 Adjustment Factor for Decision Types 73 5.5 Adjustment Factor for Time-to-Wait-for-Decision-Making 75 5.6 Adjustment Factors for Attendance-Probability-to-Communication 77 5.7 Adjustment Factor for Response Probability 78 5.8 Adjustment Factor for Probability-of-Attending-to-Communication 79 5.9 Sets of Micro-level Behavior 80 5.10 Summary of Workflows 83 5.11 Setting of Project Intensity 83 5.12 Team Experience 85 5.13 Rework and Communication Ratio 88 5.14 Summary of Simulated Results 88 5.15 Effects of Team Experience 91 5.16 Comparison of Reworking Volume 97 5.17 Comparison of Coordination 99 5.18 Comparison of Time-to-Wait-for-Decision-Making 100 Chapter 6 6.1 Global Weight Adjusted by Team Experience 117 6.2 Institutional Exceptions Distribution Policy 118 6.3 Cross-Cultural Experience Effect on Priority 119

xi

Chapter 7 7.1 American and Japanese Organization Styles 131 7.2 Matrix Strength 132 7.3 Combination Patterns of Teaming 133 7.4 Task Interdependency and Complexity 133 7.5 Simulated Results of Pooled Cases 135 7.6 Simulated Results of Sequential Cases 136 7.7 Simulated Results of Reciprocal Cases 137 7.8 Simulated Results of Intensive Cases 138 7.9 Average of Increased Duration, Work Volume, or Quality 140 7.10 Average of Increased Duration, Work Volume, or Quality (Matrix Strength) 143 7.11 Ratio of Institutional Exceptions to Total Exceptions 148 7.12 The Correlation Coefficient (r) for all cases 150 7.13 The Correlation Coefficient (r) for Pooled Workflow 151 7.14 The Correlation Coefficient (r) for Sequential Workflow 152 7.15 The Correlation Coefficient (r) for Reciprocal Workflow 152 7.16 The Correlation Coefficient (r) for Intensive Workflow 153

xii

List of Figures Chapter 2 2.1 The Onion Diagram: Manifestation of Culture at Different Levels of Depth 9 2.2 The Nature of Cultural Differences 10 2.3 Simulation Systems can Bridge the Gap between Macro and Micro Levels 27 Chapter 3 3.1 Evaluation Trajectory 33 3.2 Validation of the Reasoning Assumptions of IC-VDT 35 3.3 Research Process and Validation Steps 36 Chapter 4 4.1 A Joint Venture Team 38 Chapter 5 5.1 Inputs and Outputs of VDT Simulations 66 5.2 Adjusted Probabilities for Decision Making Policy (High Centralization) 72 5.3 Adjusted Probabilities for Decision Types 74 5.4 Adjusted Duration for Time-to-Wait-for-Decision-Making 75 5.5 Adjusted Probabilities for Attendance Probability to Communication 77 5.6 Adjusted Probabilities for Response Probability 78 5.7 A Pooled Workflow Structure 81 5.8 A Sequential Workflow Structure 81 5.9 A Reciprocal Workflow Structure 82 5.10 An Intensive Workflow Structure 82 5.11 Framework of Intellective Experiments 86 5.12 Examples of American Organization Structure Type with Intensive Complexity 87 5.13 Examples of Japanese Organization Structure Type with Intensive Complexity 87 5.14 Effects of Changes in Organization Structure Types 90 5.15 Effects of A-vs.-J Micro-Level Behavior with A style 93 5.16 Effects of A-vs.-J Micro-Level Behavior with J style 93 5.17 Effects of Product Quality Risks 94 5.18 Effects of Project Quality Risks 95 5.19 Preferred Coordination Mechanism 103

xiii

Chapter 6 6.1 Framework of the IC-VDT Model 110 6.2 Pre-Processing Model 112 6.3 A Screenshot of Pre-Processing Model 113 6.4 An Example of Micro-Level Behavior Parameters 113 6.5 A Screenshot of Selection Function 114 6.6 Institutional Exceptions Generating Mechanism 116 6.7 Institutional Exceptions Distribution Policy 118 6.8 Information Demand and Capacity 120 6.9 Framework of the IC-VDT model 123 6.10 A Screenshot of IC-VDT 125 Chapter 7 7.1 Framework of Intellective Experiment 130 7.2 Hidden Work Volume (J organization style): Case of Reciprocal Workflow 141 7.3 Hidden Work Volume (A organization style): Case of Reciprocal Workflow 141 7.4 Functional Quality Risk: Case of Reciprocal Workflow 142 7.5 Project Quality Risk: Case of Reciprocal Workflow 142 7.6 Project Duration (A org. style): Case of Reciprocal Workflow 144 7.7 Project Duration (J org. style): Case of Reciprocal Workflow 144 7.8 Functional Quality Risk (A org. style): Case of Reciprocal Workflow 145 7.9 Functional Quality Risk (J org. style): Case of Reciprocal Workflow 145 7.10 Project Quality Risk (A org. style): Case of Reciprocal Workflow 145 7.11 Project Quality Risk (J org. style): Case of Reciprocal Workflow 145 7.12 Changes in Duration for Mixed Cultural Teams 147 7.13 Team Effectiveness 155 7.14 Institutional Costs and Coordination Costs: 156 7.15 Turbulent Point and Institutional Exceptions 157 7.16 American Organization Structure Type with Interfaces 159 7.17 Japanese Organization Structure Type with Interfaces 159 7.18 Validation of the Reasoning Assumptions of IC-VDT 162 Chapter 8 8.1 Anticipated Contributions To VDT Research 177

xiv

CHAPTER ONE: INTRODUCTION

In this introductory chapter I give the purpose and motivation of my research, including

my research approach and brief outlines of the dissertation chapters.

1.1 Purpose and Motivation The 21st century may very well become known as the century of the “global world”

(McFarland et al, 1933). In the era of globalization, the boundaries between countries are

becoming negligible in many respects: for business, travel, shopping, and communication.

Interactions between countries at all levels take place much more often, facilitated by the

development of transportation and information technologies. In the construction industry,

the recent business trend toward globalization is inexorable1, and is leading to increased

cultural diversity within global construction projects. Project owners and managers are

enjoying the benefits of globalization. They can, for instance, purchase materials,

technology, and knowledge from all over the world. On the other hand, project managers

in global construction projects are now confronted with the difficulty of coordination

among sponsors, financiers, developers, planners, consultants, designers and project team

members who come from different countries. The coordination challenges are both

logistical and, frequently, cultural.

Participants in global projects work for parent companies with varying corporate

standards and management styles. The companies’ headquarters are located in different

countries, so project participants must cope with a variety of languages, business customs,

and cultures. In other words, international joint venture (IJV) teams have a great deal of

internal complexity in cultures, professions, business customs, and management styles.

Additionally, the projects are subject to varying political, economic, institutional, and

physical environments due to differences in location. Moreover, construction projects

often must be completed under great time pressure. Not only are many construction

projects highly customized to produce a one-of-a-kind facility, but in many cases, the

1 One example shows that, for Engineering News Record’s (ENR) Top 225 International Contractors, revenue from projects outside their home countries rose 20% to $139.82 billion in 2003, up from $116.52 billion in 2002 (ENR, 2004).

1

project groups also have no shared past experience working together, because teams are

assembled project-by-project.

Rapidly accelerating globalization in the construction industry has caused various

problems. For instance, Beamish finds that “international joint venture teams are

inherently risky, with failure rates approaching 50%” (Beamish, 1985; Cullen et al, 1995).

Another study shows that multi-national companies lost about US$24B during 1998 in

their private infrastructure investment activities (Irwin et al, 1997). This project was

motivated by a desire to address this high failure rate, and in particular, to show the role

that differences in culture play on project performance.

IJV studies conducted by multiple researchers reveal the difficulties of managing

mixed cultural teams. Problems such as misunderstandings and miscommunications arise

due to pre-existing differences in the participants’ values and beliefs, work standards, and

preferred business practices (e.g., Buckley and Casson, 1988; Contractor and Lorange,

1988). Also, the multiple parties comprising a joint-venture team fail to comprehend the

difference in reasoning behind their partners’ approaches. Underestimating the influence

of these cultural and institutional elements on global projects can have a potentially large

negative impact on team performance.

Institutional theory provides a broad framework, including cultural elements, to

characterize and understand the different reasoning of groups, firms, or societies. Scott

defines institutions as consisting of cultural-cognitive, normative, and regulative structures

and activities that provide stability and meaning to social behavior (Scott, 2001). This

dissertation particularly focuses on differences in “culturally-driven normative systems”

that can play an important role among participating groups in international joint ventures.

In a seminal book, Scott proposes that normative systems include both values and norms

(Scott, 2001). Values are conceptions of preferred or desirable standards of individual

behaviors. Norms specify how things should be done and define legitimate means to

pursue the value ends, specifying acceptable organizational and work practices of a group,

a firm, or a society. This dissertation presumes subgroups composing an IJV team have

and bring their own culturally-driven normative systems to a global project, causing

differences and/or coordination problems.

2

A key word used in this work is “culture.” What is culture? There is a large

literature to define culture. A common and key term in definitions of culture is the word

“shared.” Many researchers assert that culture develops from a set of shared experiences,

understandings, and meanings among members of an informal group, an organization, a

community, or a nation (Davis, 1984; Hofstede, 1991; Louis, 1985; Sathe, 1985; Schein,

1989). Most groups at any level — e.g., team, firm, society, and/or nation levels-have

their own unique culture as a result of sharing a common history and a series of common

struggles and successes. These shared experiences lead to the development of a shared set

of values and practices that are the main components of culture (Hofstede, 1991). Value-

and-practice dimensions appear to be important to understanding the reasoning behind

cultural differences.

This work extends the value-practice dimension by incorporating notions of

normative systems (Scott, 2001). Cultural values define what is right and wrong or

specify general preferences (Brown, 1976). These cultural values can be a basis for

individual behavior, for example, how people make a decision and how they communicate

with superiors/subordinates (Lane, 1992; Mankoff, 1979). Therefore, this research

extends the term “value differences” to refer to the preferred behavior patterns that people

show when making task-related and communication-related decisions in business

situations. In other words, this dissertation presumes that there are multiple patterns of

individual behaviors in international joint venture projects. In Hofstede’s research

(Hofstede, 1991), “practices” originally refers to symbols, heroes, and rituals. Since this

research focuses on the project level, it extends the meaning of “practices” to apply to

managerial and organizational norms. Thus, this research extends the meaning of

“practice differences” to include cultural norms for adopting or using specific

organization designs to manage organizations and tasks. Therefore, I propose that each

country is most likely to have its own set of typical values and practices that together

comprise “culturally-driven normative systems.”

This research focuses on two cultures — Japanese and American — as an example of

the minimum dyadic unit of cultural interaction in global construction projects. The

Japanese and American construction industries are internationally renowned as world

leaders (Levy, 1990; Flanagan, 1994). For instance, of the top 225 international

3

contractors, 42.7% are from the two countries (USA, 74; Japan, 18) (Engineering News-

Records, 2000). Additionally, Japanese and American firms have distinctive cultures and

business customs. Many researchers, for instance, have characterized distinguishing

differences between Japanese and American cultures in business situations (e.g., Nakane,

1970; Ouchi, 1981; Aoki, 1992). Their findings help in understanding the internal

consistency of Japanese vs. American social and organizational principles and their

differences from one another. However, very few cross-cultural and intercultural studies

have focused on the construction industry, even though the international construction

market was worth $106.5 billion2 per year in 2001, and has grown since then.

The first research goal of this dissertation is to observe and characterize the typical

culturally-driven normative systems of Japanese teams vs. American teams along the

values and practices dimensions through case studies (Chapter 4). Additionally, I observe

and analyze the consequence of multiple culturally-driven normative systems.

The second goal of this dissertation is to understand the impact of cultural

differences on team performance. Several researchers have tried to measure these effects

through case studies (e.g., Xiao and Proverbs, 2002; Kravis, 1984; Heston and Summers,

1996). However, researchers face many difficulties in trying to compare real projects and

isolating the pure effects of cultural differences on team performance. Factors that vary

across projects include their differing economic, political, technological, and physical

environments.

Computer simulation is growing in popularity as a methodological approach for

organizational researchers (Dooley, 2002). Simulation provides a “virtual laboratory”

where researchers can address questions about organization science with far more control

over the sets of variables that might influence outcomes. Computational laboratories

enable greater experimental variety that complements other approaches used in

organization science (Burton, 2003). Specifically, simulation models such as the Virtual

Design Team (VDT) allow researchers to ask a series of “what-if” questions (e.g., Dooley,

2002; Burton, 2003; Carley, 1995, 1996) related to the effects of variables on project

outcomes.

2 June 2001 issues of ENR magazine.

4

This work adopts the VDT model as a virtual organization laboratory for three

reasons: (1) the VDT model was built to analyze the performance of project organizations,

the appropriate unit of analysis for this research; (2) the large quantity of organizational

and individual level behavioral parameters available in the VDT model can potentially

represent culturally-driven coordination mechanisms with some fidelity; and (3) the VDT

model has been extensively validated by previous research (e.g., Cohen, 1992;

Christiansen, 1993; Thomsen et al, 1999). Furthermore, the VDT model fulfills the three

key criteria for use as a “theorem prover” to examine hypotheses (Burton & Obel, 1995)

— reality, content, and structure. Therefore, this research uses the VDT model to analyze

the effects of organizational and individual normative differences on project outcomes.

The VDT model (Jin and Levitt, 1996) extends information processing theory

(March and Simon, 1958; Galbraith, 1973, 1977) by measuring the fit between

information processing capacity and information processing demand at the level of an

individual actor. This approach is called a “neo-information processing view” (Burton

and Obel, 2004). However, VDT was not intended to model multiple cultures. This

research conducts computational experiments, called intellective experiments, in order to

explore the extent to which VDT’s information processing view of organizations can

capture the impact of cultural differences on the performance of project organizations.

Following this view, this research encodes the stochastic patterns of individual actors’

behaviors in decision making and communication driven by differing cultural values,

based on observations and literature review. In other words, we describe agents in the

VDT model representing differing cultural types. Similarly, this research models

organization structures as stochastic decision-distribution patterns driven by differing

cultural practices. I simulate the possible combinations of US vs. Japanese cultural values

and practices in each project context to predict work volume, project cost, duration, and

quality outcomes (Chapter 5).

The third goal of this dissertation is to seek a better organization design for mixed-

cultural teams. To address the limitations of the current VDT in modeling culturally-

driven organization styles and decision making behaviors, the third phase of the research

develops a prototype computational model — “Intercultural-Virtual Design Team” (IC-

VDT) — to analyze the impacts of mixed-cultural teams. IC-VDT can represent and

5

reason about multiple patterns of individual behaviors in a project. IC-VDT also address a

new concept of exceptions caused by differing organizational and work practices - called

institutional exceptions - based on field observations and micro-behaviors described in the

literature (Chapter 6). Using IC-VDT, I examine the impacts of mixed cultural teams on

team performance with the same project contexts, drawing implications about which

organization style works best for mixed cultural teams (Chapter 7).

I validate the IC-VDT model using a validation framework proposed by Thomsen

(Thomsen et al, 1999) that evaluates the model reasoning, representation, and usefulness.

The first step validates the reasoning assumptions. This research needs to correctly encode

the micro-level behavior patterns and typical practice styles of Japanese and American

teams, based on ethnographic observations (Chapter 4). The second validation step

examines the relationship between encoded micro-level behaviors and macro-

organizational behaviors, using idealized cases, called intellective experiments (Chapter 5).

The third step is to model and simulate mixed cultural team cases using IC-VDT, to

validate its reasoning for mixed cultural contexts (Chapter 6 and 7).

The claims for the contributions of this dissertation should not be exaggerated or

over-generalized. Certainly this dissertation does not cover the entire range of social

phenomena in Japanese and American life. Rather, I intend to model a small number of

key cultural attributes that contribute to team performance in the construction industry.

Additionally, this work may open a window for the use of computational modeling to

represent and study cultural differences. I believe that this work is an important first step

toward encoding cultural phenomena into organizational simulation models.

This work has been undertaken as part of the Collaboratory for Research on Global

Projects (CRGP)3, which was created to understand the institutional costs and benefits that

arise in global projects. My work will contribute to the computational modeling goals of

CRGP research.

3 CRGP is a research group headed by Professor Raymond E. Levitt (Stanford University) and supported by the National Science Foundation and industrial affiliates. For more information about the group, please see http://crgp.stanford.edu

6

CHAPTER TWO: POINTS OF DEPARTURE

This chapter reviews and outlines the three relevant research fields: cultural and

institutional theory, cross-cultural and intercultural research on Japanese and American

corporations, and organization theory. This chapter also describes the extensions needed

in order to develop methods for understanding and analyzing cultural impact on project

performance in international joint venture projects.

2.1 Cultural and Institutional Theory Research on cultural and institutional differences provides the motivation and is an initial

point of departure for this dissertation. Global project managers face difficulties in

coordinating people who come from different countries and in managing subgroups that

are headquartered in different countries. Miscommunications and misunderstanding arise

since team members have different cultural values and beliefs, and also since subgroups

have different ways to organize people and implement tasks, increasing the internal

complexity of global projects.

Cultural differences definitely play key roles in increasing the internal complexity

of global projects. What is culture? What are cultural differences? How do cultural

factors affect organizational efficiency? I need to answer these questions by defining

“culture”’ as is the first step of this research.

Institutional theory includes and covers many relevant issues such as regulative,

normative, and cultural-cognitive institutional theory (Scott, 2001). This multi-

disciplinary field, contributed to by economics, sociology, anthropology, political science

and organization science, has the potential to link among cultural theory, Japanese firm

related research and organization science in order to build a new theory about increasing

coordination cost emerging in global projects.

2.1.1 What is “Culture”?

Any cultural research needs to consider the definition of culture. At first glance, there are

a variety of possible definitions. A common and key term in definitions of culture is the

7

word “shared.” Many researchers assert that culture develops from a set of shared

experiences, understandings, and meanings among members of a group, an organization,

a community, or a nation (e.g., Davis, 1984; Hofstede, 1991; Louis, 1985; Sathe, 1985;

Schein, 1989). Groups at any level have their own unique culture as a result of sharing a

common history and a series of common struggles and successes. At a national level, for

instance, different economic classes and different generations have unique shared

experiences that lead to the formation of common sub-cultures. As a result, shared

experiences are the basic building blocks of culture. These shared experiences lead to the

development of a shared set of values and practices. Additionally, different sets of shared

values and practices exist, depending on the focal subject, since each group has its own

unique set of shared experiences.

There is a large literature on culture that describes different cultural subjects:

national culture (e.g., Hofstede, 1991; Trompenaars, 2004), founder culture (e.g., Schein,

1985), professional culture (e.g., Hofstede, 1991; Schein, 1985), and organizational and

corporate culture (e.g., O’Reilly and Pfeffer, 2000; Schwartz and Davis, 1981). Which

cultural delineation(s) do we need to consider in capturing the cultural differences that

are most salient in construction projects? In Hofstede’s research (1980), national

culture explained 50 percent of the differences in employees’ attitudes and behaviors.

National culture explained more of the differences than did professional role, age, gender,

or race (Hofstede, 1980; Schneider, 1988). Laurent found national differences more

pronounced among employees from around the world working within the same

multinational company than among employees working for different organizations in

their native lands. He assumed that mangers working for the same multinational

corporation would be more similar than their domestically employed colleagues, but

instead he found the managers maintaining and even strengthening their national cultural

differences (Laurent, 1983). Similarly, Fruchter and Townsend (2003) observed

students’ communication behaviors through the Computer Integrated

Architecture/Engineering/Construction (A/E/C) 2000/2001 class at Stanford University.

42 students, who participated in this class from different countries such as Asia, the

United States, Latin America Eastern, and Western Europe, were assigned to

geographically distributed teams to design a building, interacting by using information

8

and collaboration technologies. They found that national cultural differences showed a

stronger correlation with communication behaviors than professional differences such as

architect, engineer, or construction manager professional roles (Fruchter and Townsend,

2003). Even though it is fair to say that all of these cultural delineations are involved and

affect the groups in different ways — e.g., through the management system or through

individual preferences —, but national cultural delineation is a good primary focus for

research to understand and measure differences among participants and subgroups of IJV

projects.

Hofstede describes “culture” using an onion diagram of symbols, heroes, rituals,

and values (Hofstede, 1991, p9) (Figure2.1). He labels symbols, heroes and rituals as

practices. Values are centered in his onion diagram: they have a plus and a minus side

such as evil-vs.-good, just-vs.-unjust, and so on.

Figure 2.1: The Onion Diagram: Manifestations of Culture at Different Levels of Depth

Note: In this figure, these lines are illustrated as the layers of an onion, indicating that symbols represent the most superficial and values the deepest manifestations of culture, with heroes and rituals in between (Hofstede, 1991).

Hofstede’s research suggests that the practices-values balance is shifted in

conjunction with changes in the size of groups (Figure 2.2). He describes that at the

individual level, cultural differences reside mostly in values, less in practices. However,

at the organization level, cultural differences reside mostly in practices and less in values.

Therefore, at the project level, the primary subject of this research, cultural differences

reside in some middle ground between the two levels (Figure 2.2). Additionally, many

comparative studies of culture have used practice and value dimensions to interpret

managerial differences among countries (e.g. House et al, 2004; Botti, 1995), but they use

9

“practice” meaning organizational and business forms and customs. Thus, this research

starts to view cultural differences from these two points of view: practices and values.

Organizational Level (Molecules)

Organizational Culture Practice Differences

Project Level

Individual Level (Atoms)

Figure 2.2: The Nature of Cul

Balance Proposed by Hofstede

Note: This illustrates that at tvalues, less in practices. At thpractices, less in values. Thidifferences is dependent on thfocuses on the project-organiequally involved. Additionalare attributed to national, fou

2.1.2 Cultural Values

Recent attempts to develop the

values as the important elemen

Efforts were made to identify

which different countries can

1991; Trompenaar, 1993; Tria

conceptions of the preferred a

which existing structures or be

other words, values are core to

feel, think, or act in a given sit

research indicates that cultural

dimensions: power distance, in

Founder’s Culture

D

s

z

n

a

b

n

National Culture

tural Differences; Adapted from t

(Hofstede, 1991, p.182)

he individual level cultural differe organization level, cultural diff implies that the value-practice be size of group (Hofstede, 1991)ation level, we assert that the va

ly, this research assumes that cultder, professional, and organizati

oretical frameworks for culture h

ts of culture that impact organiza

nd classify the cultural values or

e compared and contrasted (e.g.,

ndis, 1982; Ronen and Shenkar, 1

d desirable, together with the co

havior can be compared and asse

“mental programming,” which g

uation and context (Hofstede, 19

values can be represented by the

dividualism vs. collectivism, ma

Value ifferences

ProfessionalCulture

he Values-Practices

ences reside mostly in erences reside mostly in alance in cultural . Since this research lue-practice elements are ural values and practices onal cultures.

ave pointed at cultural

tions (Hofstede, 1980).

patterns of values, by

Glenn, 1981; Hofstede,

985). Values are

nstruction of standards to

ssed (Scott, 2001). In

uides people in how to

91). Hofstede’s (1991)

following five

sculinity vs. feminity,

10

uncertainty avoidance, and short- vs. long- term orientation. Understanding the nature of

these dimensions can help us predict how individual team members tend to behave while

communicating and making decisions.

Power distance (PDI): Power distance is the extent to which the less powerful

members of organizations and institutions accept and expect that power is distributed

unequally. This inequality is defined from lower tiers of the power hierarchy (those with

less power). It suggests that a society’s level of inequality is endorsed by the followers as

much as by the leaders.

Individualism vs. Collectivism (IDV): Individualism refers to the degree to

which individuals are integrated into groups. On the individualist side, Hofstede

describes societies in which the ties between individuals are loose: everyone is expected

to look after himself or herself and his or her immediate family. On the collectivist side,

he describes societies in which people, from birth, are integrated into strong, cohesive, in-

groups, often extended families which offer protection in exchange for unquestioning

loyalty.

Masculinity vs. Femininity (MAS): Masculinity versus femininity refers to the

distribution of roles between the genders. The IBM studies conducted by Hofstede

(1991) revealed that women’s values differ less amongst societies than do men’s values

amongst different countries. These values vary along a dimension, from assertive,

competitive, and maximally different from women’s values on one extreme, to modest,

caring, and similar to women’s values on the other. The assertive pole has been called

‘masculine’ and the modest, caring pole ‘feminine.’ The women in feminine countries

have the same modest, caring values as the men; in the masculine countries they are

somewhat assertive and competitive, but not as much as the men, so that masculine

countries show a large gap between men’s values and women’s values.

Uncertainty Avoidance (UAI): Uncertainty avoidance deals with a society’s

tolerance for uncertainty and ambiguity; it ultimately refers to man’s search for truth. It

indicates the extent to which a culture programs its members to feel comfortable in

unstructured situations. Unstructured situations are novel, unknown, surprising, and

different from the usual. Uncertainty-avoiding cultures try to minimize the possibility of

such situations by using strict laws and rules, and safety and security measures. On the

11

philosophical and religious level they believe in absolute Truth: “There can only be one

Truth and we have it” (Hofstede, 1991). People in uncertainty avoiding countries are also

more emotional and motivated by inner nervous energy. The opposite type, uncertainty-

accepting cultures, is more tolerant of different opinions. They try to have as few rules as

possible, and on the philosophical and religious level they are relativist. People within

these cultures are more phlegmatic and contemplative, and their countrymen do not

expect them to express emotions.

Long-Term vs. Short-Term Orientation (LTO): People with a long-term

orientation are characterized by persistence, thrift, and a sense of shame. They order

relationships by status and observe this order. In contrast, people with a short-term

orientation possess personal steadiness and stability. They care about protecting "face,”

respecting tradition, and reciprocating greetings, favors, and gifts.

Table 2.1: Raw Score along Each Dimension for Japan and the USA (Hofstede, 1991)

Japan USA Score Gap

PDI 54 40 14

MAS 95 62 33

IDV 46 91 45

UAI 92 46 46

LTO 80 29 51

Note: The score ranges from 0 to 100. The closer to 100, the more likely that the country will have that value strongly. For instance, the MAS index of Japan, at 95, indicates its strong MAS characteristics.

As Table 2.1 shows, there are large score gaps between Japanese and Americans

in the MAS, IDV, and UAI dimensions (Hofstede, 1991). These differences imply that

the two nations have differing workplace values, and that large differences in these values

can potentially cause conflicts or misunderstandings when working together. In

particular, in the case of the MAS dimension, the two countries tend to handle conflicts in

different ways (Hofstede, 1991, p92). For instance, in high masculinity countries,

organizations are most likely to solve problems using a “let the best man win” approach,

12

rather than through compromise and negotiation. These different ways of handling

conflict may cause misunderstandings or errors in judgment that stem from differences in

cultural values. Thus, the larger the gap between the two countries, the more frequently

misunderstandings occur.

2.1.3 Cultural Practices

In Hofstede’s research (Hofstede, 1991), “practices” originally refer to symbols, heroes,

and rituals. Since this research focuses on the project level, it extends the meaning of

“practices” to apply to managerial and organizational practices. Frequently, cross-

national studies attribute differences in management or organization practices to cultural

differences (e.g., Howard et al, 1983; Cavusgil and Yavas, 1984; Vertinsky, 1990, etc).

Although the research does not provide sufficient explanations of how culture causes

such differences (Lachman et al, 1994), it does provide one important aspect of cultural

differences: organizational practices. In organization research, “practices” have long

been the primary focus of attention, although cultural forms have traditionally been

dismissed as esoteric (Martin, 2002). One main reason for this focus is that questions

concerning the most efficient organization and how to design an effective organization

are the traditional agenda in organization studies. Burton and Obel (2004) describe

organization design as a normative science that concerns itself with how an organization

is put together, who does what, and who talks with whom. In this sense, “practices” can

be linked to cultural norms that suggest how one culture organizes and manages people

and resources in order to achieve goals and tasks. Therefore, this research extends the

meaning of “practices” to include cultural norms for adopting or practicing specific

project management styles and organization structures.

According to Burton and Obel (2004), organization design can be identified by

five attributes: specification of configuration (depth of hierarchy), level of centralization,

level of formalization, incentives, and coordination and control.

The configuration can be thought of as the organization chart. The configuration

specifies the general principles for dividing work, breaking tasks into subtasks and

coordinating activities. Many organizational configurations have been proposed and

discussed from time to time. For instance, Mintzberg proposed that there are five

13

archetypal configurations of organizations: the simple structure, the machine bureaucracy,

the professional bureaucracy, the divisionalized form, and the adhocracy (Mintzberg,

1980).

The simple structure: A flat hierarchy and a singular head for control and

decision making.

The machine bureaucracy: highly routine operating tasks, very formalized

rules and regulations, tasks grouped into functional departments, centralized

authority, decision making follows the chain of command and an elaborate

administrative structure with sharp distinction between line and staff.

The professional bureaucracy: Highly skilled professionals, high

complexity, decentralization, and internal professional standards.

The divisionalized form: self-contained unit groupings into somewhat

autonomous units coordinated by a headquarters unit (product, customer, or

geographical grouping, including multinational).

The adhocracy: high horizontal differentiation, low vertical differentiation,

low formalization, decentralization, and great flexibility and responsiveness.

These dimensions are further explained and discussed in Chapter 5. Interestingly,

Hofstede links Mintzberg’s typology of organizations to cultural factors along two

dimensions: power distance and uncertainty avoidance. Hofstede’s research (Hofstede,

1991) provides a potential interpretation of why a certain organizational configuration

has been selected and/or emerged in a firm or a country.

Level of Centralization describes how much power a project manger (PM)

delegates to subordinates. The less a PM delegates, the more centralized the team will be.

Centralization reflects whether decisions are made by a senior PM or decentralized to

team members or a sub team leader. High project centralization indicates that almost all

decisions are made by a PM. With decentralization, a sub team leader or team members

tend to make their own decisions.

Level of Formalization is a measure of how structured tasks and communications

are in an organization. With respect to communication, high formalization refers to a

tendency for communication to occur using formal meetings. With low formalization,

communications occur ad-hoc between team members. For many organizations, it is

14

efficient to obtain standardized behavior from the members of the organization. This

standardization can lead to low cost, high product quality, and generally efficient

operations. Formalization is one way to obtain such standardized behavior and thus is an

efficient means to increase coordination and control (Burton and Obel, 2004).

The incentive system describes how individuals and their activities are evaluated

and compensated. Abraham Maslow proposed five basic human needs which motivate

people according to a hierarchy of needs (Maslow, 1943): basic needs, security needs,

belonging needs, esteem needs, and self-actualization needs. His work implies that

monetary award, job security, the reputation of the organization in the society, and

promotion systems all need to be taken into consideration in considering the incentive

system. Nakane (1970), for instance, argues that Japanese employees tend to be

motivated to increase their rank over their salary.

There are two sides to considering coordination and control. One is to insure

that enough relevant information is available at the right time to be able to make the right

decisions. The other is to make sure that the right decisions are made. This includes the

development and standard of rules and procedures, in the form of meetings, quality,

safety, accounting, and information technology systems for communication.

2.1.4 Institutional Theory

Institutional theory discusses broader issues concerning multi-national corporations.

Studying and understanding the institutional aspects enable researchers like myself to see

global issues from multiple points of view and deepen the arguments. Scott (2001)

proposed three institutional pillars: regulative, normative, and cognitive pillars.

Institutions consist of cultural-cognitive, normative, and regulative structures and

activities that provide stability and meaning to social behavior. Institutions are

transported by various carriers — cultures, structures, and routines — and they operate at

multiple levels of jurisdiction (Scott, 1995). In particular, this dissertation emphasizes

the normative aspects of institutions. A normative system gives priority to moral beliefs

and internalized obligations as the basis for social meaning and social order. In this

conception, behavior is guided not primarily by self-interest and expedience, but by an

awareness of one’s role in a social situation and a concern to behave appropriately, in

15

accordance with others’ expectations and internalized standards of conduct (Table 2.2).

Parsons (Parsons, 1960) developed his cultural-institutional approach by examining the

ways in which the value system of the organization was legitimated by its connection to

the wider societal institutional norms and values.

Table 2.2: Institutional Elements and Carriers

Elements

Carriers Regulative Normative Cultural-Cognitive

Cultures rules, laws values, expectations categories, typifications

Structures governance systems,

power systems

regimes, authority

systems

structural isomorphism,

identities

Routines compliance,

obedience

conformance of duty performance programs,

scripts

Note: Adapted from W.Richard Scott. (The institutional construction of organizations, Sage, Thousand Oaks, CA, 1995, Table 1.1)

In a seminal book (Scott, 2001), he further discussed the dynamics of institutions.

In particular, this section discusses three relevant aspects to this research: isomorphism,

cultural persistence vs. adaptation, and the stability of the value system.

Isomorphism: Can we assume that Japanese and American teams tend to have a

certain type of organization architecture? Institutional theory can answer and support this

question to a certain degree. The principle of isomorphism, first applied by Hawley

(1950, 1968), suggested that units that are subjected to the same environmental

conditions, and units that interact frequently acquire a similar form of organization.

Moreover, DiMaggio and Powell (1983) stressed that institutional mechanisms, such as

coercive, normative, and mimetic systems, tend to make organizations more alike without

necessarily making them more efficient. This implies that construction firms in each

country, doing business under the same institutional, economic, and cultural

environments, become more likely to have a similar organization design, and hence,

practice style.

Aoki (1992) suggests that typical Japanese business organizations tend to cluster

16

toward one end of a prototypical spectrum of values and practices, (relatively)

independently of the individual organizational environment. Their American

counterparts tend to cluster near the other end of this spectrum, also independently of

their environment. Other researchers (e.g., Nakane, 1970; Ouchi, 1981; Nonaka, 1995)

attempted to categorize the typical organizational styles of Japanese and American firms,

and to explain their differences from a perspective of cultural values, norms and beliefs.

This implies that there is large possibility that each country tends to foster its own set of

organizational and managerial practices.

Furthermore, the isomorphism theory and these empirical findings cast a question

why a certain organization form has been selected and fostered over years in a country.

In general, scholars chose to focus on one, or at most a few, factor(s) they consider

dominant. There are three prototypical views, which may not be entirely exclusive.

The culturalist view: Cultural values, beliefs, and norms can lead to the

development of “acceptable practices” within a culture and an organization

(e.g., Abegglen, 1958).

The universalist view: In any given environment, there is a specific

organizational design that is most efficient. Therefore, similar

organizational designs tend to be selected in Japan and U.S. for the same

environment (e.g., Koike, 1975).

The institutionalist view: various institutional processes — i.e., regulative,

normative and cultural-cognitive systems — combine with other forces —

i.e., technical and economical pressures — to shape an organizational

design in a country for years or decades (e.g., Scott, 2001).

I emphasize that these views are only prototypical, in that scholars generally

combine more than one view, either explicitly or implicitly. To answer this question is

not the primal purpose of this dissertation. However, exploring the linkage between

cultural values and cultural practices through computational experiments can potentially

contribute to this question.

Cultural persistence vs. adaptation: My research collects data from joint-

venture projects around the San Francisco Bay area, comparing Japanese and American

teams in the United States. The key benefit of this methodology is to control for

17

exogenous environmental factors such as regulative institutions and physical/economic

complexity. Under this research scenario, it is worthwhile to discuss how organizational

behavior changes in a foreign environment. Will subsidiary offices headquartered in

another country maintain their original cultural practices and values? Or can they quickly

adapt to local practices? Scott (2001) argued that two opposing forces exist for this issue:

persistence vs. adaptation. Simon, for example, emphasized a cognitive pattern: “The

activity created stimuli that directed attention toward its continuance and completion”

(Simon, 1945, p106). Researchers (e.g., Simon, 1947; Hannan and Freeman, 1984, 1989)

argue that persistence implies that change is assumed to be both difficult and dangerous

for organizations. Other researchers, such as Zucker (1988), suggest that a tendency

toward disorganization and decay in the social system is the more normal condition.

Things – structures, rules, routines – tend to fall apart.

At the project level, structure is relatively flexible, because the project manager is

generally authorized to design the structure of the group. Japanese project managers, for

instance, may select a typical American structure, following a “When in Rome, do as the

Romans do” strategy. However, they may feel persistent pressure from their Japanese

employees or the company headquarters in Japan. Foreigner groups experience the

dilemma of how to maintain legitimacy in both local and home cultures simultaneously.

Kostova and Kendall (2002) argue that subsidiaries of global companies are substantially

exposed to two different institutional pressures, the local and home country, which they

term dual institutional pressures. Investigations in this field are quite new so it will be

interesting to discover how each subsidiary group balances the two pressures.

Stability of Value Systems: I have discussed how cultural values and beliefs can

be significantly attributed to the particular management practices of firms. Can we

assume that cultural values are relatively stable? As a result of globalization in the past

decades, people have often been exposed to other cultures. A first glance at

contemporary Japan reveals Western cultural influences through music, food, language,

and Hollywood movies. People, particularly the younger generations, may have different

values and beliefs. Lincoln and Kalleberg (1990), for instance, pointed out that the

younger generations of Japan have come to spend more time on leisure rather than

working for their corporation. Odaka (1975) and Sengoku (1985) also argued that the

18

levels of loyalty and commitment of younger generations to their corporation have fallen.

This implies that cultural values are relatively unstable, especially between generations.

On the other hand, Nakane (1970) emphasized the persistence of social structure,

underpinned by core cultural values. She drew upon examples showing that the Japanese

younger generation soon begins to follow the traditional order, once they are employed,

despite their earlier tendencies to infringe upon the rules of order.

Lachman (1994) nicely explained the above arguments by addressing the concepts

of core values and periphery values. The core values tend to maintain continuity, because

they are more stable and resistant to change. Their social control effects are more

enduring. Therefore, the core values are defined as the core of the stabilizing

mechanisms of the social system. On the other hand, periphery values are less stable and

enduring, because members of society may manifest different levels of attachment to

them, or may even disregard them. Thus, core values are the high priority values central

to a social, cultural or individual’s value system. They are important in regulating social

behavior, and tend to endure. Periphery values refer to values of low priority, low

consensus, high divergence, high ambiguity, and less importance for social control

(Lachman, 1994).

This research sees the national cultural dimensions proposed by Hofstede as core

values, which are relatively stable and enduring. Of course, we cannot deny the

possibility of the core values changing as a consequence of innovative or diffused

periphery values. This research does not cover values’ diffusion level, the factors

influencing their spread, and the dynamics of core and periphery values. However, the

proposed model will have the potential to analyze, using simulation methodology, what

would happen if core values were to shift in a certain direction.

2.2 Cross-Cultural and Intercultural Research on Japanese and

American Corporations The emergence of a globally competitive market has led to a greater demand for

understanding different cultural values and practices. In particular, the emergence of

Japanese firms’ competitiveness in international markets stimulates many scholars to

study not only distinctive Japanese business practices, but also unique Japanese cultural

19

values. I categorize previous research related to Japanese firms into three types using

Hart’s terminologies (1998). Hart (1998) proposes that there are three types of cultural

studies: mono-cultural studies, cross-cultural (comparative) studies, and intercultural

studies. Mono-cultural or single culture studies are common in anthropology and

sociology. Cross-cultural (comparative) studies are studies that compare the

characteristics of two or more cultures — i.e., Hofstede (1991) proposes the five value

dimensions to compare the characteristics of 53 countries. Intercultural studies are

studies that focus on the interaction two or more cultures and answer the main question of

what happens when of two or more cultures interact (at the interpersonal level, group-

level or national level). This section discusses cross-cultural and intercultural studies,

which are key aspects to this research.

2.2.1 Cross-Cultural Research

Cross-cultural (comparative) studies (e.g., Abegglen, 1958; Nakane, 1970) have looked

into distinguishing differences between Japanese and American firms. Major

contributions have been made from fields as varied as anthropology, economics,

sociology, organization science, and business. These researchers collected data from the

automobile, utility, information technology, mechanical, and banking industries, studying

trends of the past three decades. I summarize some key results here in order to make sure

their findings are similar to those of the construction industry, and to understand the

consistency of Japanese and American social and organizational principles. The typical

approach taken is to line up distinguishing organizational styles or coordination styles

with empirical data, and give interpretations from cultural or information processing

points of view of why Japanese and American firms adopt certain practices.

I categorize their findings into five attributes relevant to organization types:

organizational configuration, centralization, formalization, incentives, and coordination

and control.

The organizational configuration includes horizontal and vertical differentiation.

The common tendency is that the American type (Type-A) is most likely to have high

functional specialization, which is associated with horizontal differences in departments

and sections, while the Japanese type (Type-J) shows less specialization. The Type-J

20

firms tend to have a job rotation system whose goal is for employees to acquire

knowledge in multiple fields so that employees can handle a variety of contingent

situations (Cole, 1979; Fujimoto, 2004; Aoki, 1994). On the other hand, Type-A firms

are apt to divide the project into a number of separate tasks, and then assign them to

specialists. Another distinguishing factor appears in their vertical differentiation. Type-J

firms tend to have high differentiation in the vertical hierarchy, because of a strict

ranking and seniority system. Therefore, Type-A firms show higher differentiation in the

horizontal hierarchy and less differentiation in the vertical hierarchy compared to Type-J

firms.

Centralization refers to the degree to which formal authority to make discretionary

choices is concentrated in an individual, unit, or level (Burton and Obel, 2004). Type-J

firms tend to have a consensual decision-making system, implying a decentralized system.

On the other hand, the expectation that subordinates employees report often to their

supervisors implies a centralized system. Interestingly, Nakane (1970) relates that

“senpai-kohai (先輩―後輩)” (senior-younger / boss-subordinate) relationships are more

important than the real decision makers’ or workers’ skills. This implies that it is better

to think of two types of centralization: general authority and real decision making. While

a boss keep general authority for maintaining “senpai-kohai’ relationships, he or she

allows decentralized decision-making to maintain the legitimacy of the group-oriented

culture. Lincoln and Kalleberg (1990) proposed a categorization of formal centralization

(general authority) and de facto centralization (decision making authority). This duality

of centralization in Japanese organizations makes the Japanese structure very unique, and

may enable it to maintain legitimacy for Hofstede’s two dimensions of high collectivist

and high-medium power distance (Table 2.1).

Formalization refers to the rules in an organization that can insure standardized

behavior from its members, leading to low cost, high product quality, and generally

efficient operations (Burton and Obel, 2004). Lincoln and Kalleberg (1990) explained

that Type-J firms need intensive clerical work because of the extremely cumbersome

Japanese written language, which still requires transcription by hand. However, this

interpretation is not yet clear, because it may not be a language issue. Ouchi (1981)

argued that Type-A firms use many formal documents, reflecting a contract-oriented

21

culture. Thus, his work implies that Type-A firms are apt to have high formalization. He

also argued that informal communications are considered important in Type-J firms,

implying less formalization. However, Type-J firms use a system called “Ringi (稟議),”

where documents are circulated for consensus, indicating that the process of decision-

making is standardized and rule-based. For instance, documents contain a standardized

area reserved for stamps, indicating a list of roles that are required to view and stamp the

document, such as a general manager, a project manager, an engineer and so on. Highly

standardized decision-making may be necessary for Japanese company, because it can

avoid certain conflicts about who should involved in a decision. Thus, both countries

may have a medium to high formalization level.

The incentive system includes job security and personnel advancement. Type-J

firms generally have a lifetime employment guarantee, various training and educational

programs, and a ranking (seniority) system, which leads to a slow evaluation and

promotion process. Type-A firms are prone to short term employment, and use a rapid

evaluation and promotion system. The lifetime employment system assumes that most of

the employees will work for the same company for their entire work life.

Coordination and control systems refer to the coordination mechanisms and

principles used to manage people and tasks. Type-J firms use “integration,” while Type-

A firms use “differentiation” or “specialization.” This implies that Type-A firms are

more likely to divide the project into a number of separate tasks and assign them to

specialists. Type-J firms are apt to take the integrated teamwork strategy, on the

principle that all tasks should be done by groups of employees with multiple skills.

2.2.2 Intercultural Research

Since the 1960s, markets have become more globalized (Hofstede, 1991). It is thus more

important to study and discuss intercultural interactions. Intercultural studies focus on

interactions of two or more cultures and answer the main question of what happens when

two or more cultures interact at the interpersonal level, group-level, firm-level or national

level (Hart, 1998). How do intercultural interactions affect the ways in which project

managers operate within organizations? What kind of organization style is best for

organizing multicultural teams? Adler (1997) argues that multicultural teams have the

22

potential to produce either negative or positive outcomes compared to mono-cultural

teams.

Research on IJV projects reports misunderstandings, miscommunications and

difficulties in reaching a consensus among participants as a negative outcome (e.g.,

Beamish, 1985; Cullen et al, 1995). For instance, American negotiators, who have to

face Japanese people accustomed to their conventional ways of thinking across

negotiating tables, are often confused and puzzled (Graham and Sano, 1984). American

negotiators tend to feel that Japanese never seem to get to the core of problems or even

onto the topic (Japanese think that it is imperative for both sides to reach mutual

understanding about the standpoints of all people involved). Young (1982) explained this

by generalizing Japanese vs. American negotiation approaches: Japanese tend to do

“relations first, talk later” approach, while Americans want to “deal first, talk later.”

There are possible reasons behind this scenario such as differing cultural values, norms,

beliefs, and business customs among the two cultures. At least, we can say that high

tensions between the two cultures often arise, negatively affecting project performance.

In addition, if one judges the other culture as bad, team members can exaggerate

misunderstandings and/or miscommunications, and fall into polarized situations such as

dislike, mistrust, offensive, racist, or ethnocentric attitudes and behaviors (Hofstede,

1991; Adler, 1997), called “naïve realism (Robinson, 1997)4.”

Adler (1997) argued that intercultural interactions can produce positive outcomes

when organizations want to expand their perspectives, their approaches, and their ranges

of ideas. Ziller (1979) also concluded that potential advantages of intercultural

interactions include enhanced creativity, flexibility, and problem-solving skills,

especially on complex problems involving many qualitative factors. A good example is

the Total Quality Management (TQM) system that has been fostered by long term

interactions between American and Japanese firms. As far back as the 1960s and the

1970s, Toyota learned the Quality Control (QC) system from American firms, and then

4 Robinson (1997) found that partisans tend to distort their evaluation of information in ways that cause further polarization and conflicts, and undermine negotiation, rather than, as one might initially expect, helping the two sides move closer together. When partisans fall into highly polarized situations, they generally show three tendencies: partisans will a) exaggerate their opposition’s extremism, b) perceive their opposition to be ideologically biased, and c) overestimate the true magnitude of their conflicts. These tendencies are called “naïve realism.”

23

adopted it in Japan, developing the Total Quality Control (TQC) system. American firms

researched TQC as a part of distinguished Japanese firms’ methods — i.e., Kaizen and

Kanban systems —, and then modified it to the TQM system. Therefore, intercultural

interactions can produce positive outcomes, both in the short and long term.

Adler (1997) also discusses linkages between organization styles and project work

stages. Adler suggests a “divergence” style is common during the early stages of a project,

because a project team creates ways of defining its objectives, gathering and analyzing

information and developing alternative forms of action, and because a divergence style

enhances creativity and innovation (Adler, 1997). On the other hand, a convergence

style becomes important during the final stages of projects, since teams need to agree, or

converge, on which decisions and actions to take. In addition, Adler argues that cultural

diversity makes work processes easier during the earlier stages, because project teams

can employ a divergence style to take advantage of differences in experience, expertise

and perspective (Adler, 1991). On the other hand, cultural diversity makes work

processes difficult during the final stages, since project teams need to converge and

integrate all ideas and options. In this regard, global projects may have more exceptions

during the final stages, or at intermediate stages requiring convergence styles.

There are significant differences between the two social and organizational

structures, indicating differences in underlying values and beliefs. Additionally, when

the two cultures interact with one another, cultural differences in values and practices

play out either negatively or positively, depending on project context, project

requirements, and work stage. I believe the exploration of these values and beliefs in

terms of their effects on organizational structure is interesting and fascinating to

researchers of business, as well as social and organizational science. Thus, cross-cultural

(comparative) and intercultural research between Japan and America is fertile ground for

the development of a theory of 21st century global social, organizational, and institutional

structures.

2.3 Organization Theory This research seeks an answer to how to design effective organizations for global

construction projects. Organization theory provides us many insights to understand how

24

to design organizations to achieve goals effectively. Etzioni (1964) defined organization

as “social units (or human groupings) deliberately constructed and reconstructed to seek

specific goals.” Organizations are designed to coordinate resources, including human

resources, to accomplish their goals. Coordination is the central purpose of existence for

organizations. Without coordination, we would not have an organization (Burton and

Obel, 2004). Coordination activities require the exchange, processing, or sharing of

information, in order to minimize risks or uncertainty. Since Weber’s fundamental work

in the early 1900s (Weber, 1924), many researchers have seen organization as

information processing systems (Simon, 1945; March and Simon, 1958; Galbraith, 1973,

1977; Stinchcombe, 1990). In this view, an organization is an information-processing

and communication system, structured to achieve a specific set of tasks, and comprised of

limited capacity, “boundedly rational” information processors (individuals or sub-teams).

The information processing view provides a framework to design effective

organizations. The basic design problem is to design an appropriate organization that

matches the demand for information processing with the appropriate information

processing capacity. Galbraith (1973, 1974) presented the organizational design problem

as an information-processing problem: “The greater the uncertainty of the task, the

greater the amount of information that has to be processed between decision makers.”

The task uncertainty can arise from the technology and environment (Thompson, 1967)

as well as other resources (Burton and Obel, 2004). High task uncertainty creates a high

demand for information processing, potentially overwhelming the information processing

capacity of an organization. Organizations can either reduce their need for information

processing or increase their capacity to process information (Galbraith, 1974). For

instance, integrated CAD-CAM systems can increase the information processing capacity

of construction firms. And slack time or budgetary targets can reduce the demand for

information processing.

Levitt and other Stanford researchers (e.g., Jin and Levitt, 1996; Thomsen, 1998)

have extended the information processing view by measuring the fit between the

information processing capacity and the information processing demand at the level of an

individual actor. This micro view of the information processing approach to

organizational design is called “neo-information processing” (Burton and Obel, 2004).

25

From a neo-information processing view, individual actors have different information

processing skills, different communication and decision behaviors as well as professions

for different communication media. These individual actors’ properties combined with

team building and organizational structure determine the information processing capacity

of an organization, affecting organizational performance.

Computational models, such as the Virtual Design Team (VDT) model (Levitt et

al, 1994: Jin and Levitt 1996) and OrgCon (Burton and Obel, 2004), have gained

popularity as managerial tools to design organizations for complex projects or

corporations in particular industries. These computational models have potential, not

only as managerial tools, but also as experimentation platforms to test relevant

hypotheses.

Thomsen et al (1999) argued that simulation systems such as the VDT model can

bridge the gap between theory and experience at a micro-5 and macro-level6 (Figure2.3).

Simulation of micro-behaviors and their interaction generates macro-predictions that can

be tested against both predictions of macro theory and macro organization experience. In

other words, the VDT model provides a meso-level7 analysis tool. For instance,

researchers can address cultural cognitive and social psychological behavior as “micro-

theories”. If researchers can appropriately encode the organizational behavior of

individual actors, based on cultural cognitive and social psychological theories, the

generated macro-behavior of the organization generated through simulation should match

predictions of organization macro-theory and macro-experience.

5 Micro level refers to organizational behavior theory that seeks answers to important questions about individuals and small groups and draws primarily from psychology, including theories of cognition and decision making, emotions, personality, groups, and social and cultural psychology. 6 Macro level refers to organizational contingency theory that seeks to understand how organizations come to be structured the way they are, how they are related to each other, and how their structures and relationships change over time, as well the effects of these structures and relationships on individual members and organizational performance, drawing primarily from sociology and economics. 7 Micro-level and macro-level organization fields come together in what has come to be known as meso-level organizational works. This cross-level work bridges the individual and organizational levels by analyzing how organization-level phenomena (such as organization design, compensation systems, culture and identity) shape individual and group behavior and how individual actions in turn shape organizational processes and outcomes.

26

Organization Macro-theory

Organization Macro-experience

Organization Micro-theory

Organization Micro-experience

Simulation micro-behavior

Emergent simulation macro-behavior Simulation System

Figure 2.3: Simulation System can Bridge the Gap between Macro and Micro Levels

Note: Simulation models such as the VDT model can mediate between organization micro-theory and micro-experience and organization macro-theory and macro-experience (Thomsen et al, 1999).

The VDT model has broken new ground, scientifically and technically, in its

accurate predictions of schedule, cost, and process quality performance. However, the

VDT model can only represent technically complex, but “mono-cultural,” engineering

teams. In other words, boundedly rational agents in the VDT model do not represent

differing values, cultural norms, or entrenched work practices. So the current VDT

model can not help managers assess and mitigate cultural complexity in global projects.

How can we model multiple agents who have different cultural values and practices? In

other words, the current VDT has been developed and validated for mono-cultural studies.

Can we capture cross-cultural and intercultural phenomena based on the information

processing abstraction in VDT? What elements do we need to consider for extending

VDT for cross-cultural and intercultural studies? These questions are the fundamental

starting points of this research.

27

CHAPTER THREE: OBJECTIVES AND APPROACH

This chapter describes the objectives and approach of this research. The main objectives

of this research are to understand, analyze, and model effects of cultural differences on

team performance, through development of a new computational model that incorporates

cultural variables observed in international joint-venture (IJV) teams composed of

Japanese and American firms.

3.1 Research Objectives and Research Questions Research Objectives:

The research objectives of this project are as follows:

Identify and understand the micro-behaviors associated with cultural values

and practices of the two countries, from an information processing point of

view.

Identify and understand quantitative and qualitative relationships among

value differences, practice differences, and organizational efficiency.

Explore effective organization designs for mixed cultural teams by extending

the VDT model.

Validate reasoning of the proposed model (the extended version of the VDT

model).

Research questions:

There are three main questions this research aims to answer:

What kinds of cultural differences are at play in global projects, and

which ones are critical to team performance?

- What are the distinctive culturally-driven behavior patterns in global

projects? Can we explain culturally-driven behavior by the cultural

values dimensions proposed by Hofstede?

28

- What are the distinctive culturally-driven practices in global projects?

Are observed practices in the construction industry consistent with

empirical findings from the literature survey?

How much do cultural values and practices affect team performance?

- How can we appropriately encode value-and-practice related

parameters based on observations?

- How much do value differences affect team performance? Which

team outcome is the most influenced by changes in individual

behavior patterns?

- How much do practice differences affect team performance? Which

team outcome is the most influenced by changes in organizational

styles?

What is a better organization design for global projects?

- How can we model multiple culturally-driven normative systems on

global projects, and validate the proposed model to predict

consequences?

- How much do mixed cultural teams influence team performance?

- What leadership style works best to lead and organize a mixed cultural

team?

- How strong are relationships between independent and dependent

variables of the proposed model?

3.2 Research Approach Three building blocks provide the basis for this dissertation: cultural cognitive sociology,

organization science, and Japanese firm-related research (cross-cultural and intercultural

research).

Research on national cultural differences provides the motivation and is an initial

point of departure for this dissertation. Miscommunications and misunderstandings arise

among people and groups who come from different countries. In other words, global

project managers face difficulties in coordinating people who come from different

countries and in managing subgroups that are headquartered in different countries. These

29

coordination problems illustrate that internal complexity is increased in international

joint-venture projects in comparison to mono-cultural teams. Cultural differences

definitely play key roles in increasing the internal complexity of global projects. What is

culture? What are cultural differences? I need to answer these questions by defining

“culture”’ as is the first step of this research.

This research focuses on two cultures: American and Japanese. There are clear

differences between the two countries in cultural values and preferred organizational

designs. I believe the exploration of these values and beliefs in terms of their effects on

organizational design is interesting and fascinating to researchers of business, as well as

social and organizational science. Additionally, the two cultures represent the minimum

dyadic unit of cultural interactions so that there is the potential to apply findings drawn

from this study to other cultures.

How do cultural factors affect organizational efficiency? In order to answer this

question, this research needs to link cultural and institution theory and organization

science. Cultural cognitive research provides theoretical frameworks about people’s

behavior patterns, as micro-level organization theory. Cultural normative research

accommodates people’s perceptions and preferences of work practices in their work

groups, providing reasoning about organization design. Organization science describes

organizational behavior as a macro-level theory. Institutional theory includes and covers

many relevant issues such as regulative, normative, and cultural-cognitive institutional

theory (Scott, 2001). This multi-disciplinary field, contributed to by economics,

sociology, anthropology, political science and organization science, has the potential to

explain and build a new theory about institutional conflict issues emerging in global

projects.

Organization science is needed to predict team performance. Cultural factors that

are involved in projects teams are captured, based on organization theory. For predicting

team performance at the project level, a computational model such as VDT (Levitt et al,

1999) provides both a useful laboratory for experimentation and a potential platform to

discover better project-organization designs for global projects. The VDT model is based

on organization theories such as the information processing view of contingency theory

(Galbraith 1974, 1977). This research is categorized as an extension of contingency

30

theory: a micro-contingency theory of information processing demand and capacity based

on cultural differences.

3.3 Research Steps This research is composed of four phases: data collection, intellective experiments for

single-cultural cases, extension of VDT, and intellective experiments for mixed cultural

cases. Phase 1, 2, 3, and 4 are discussed in the following sections respectively:

Phase 1: Chapter 4: Data collection to characterize the culturally-driven

normative systems of Japanese vs. American teams along with value-

practice dimensions

Phase 2: Chapter 5: Intellective experiments to examine the impacts of

cultural differences and to validate the encoded cultural parameters

Phase 3: Chapter 6: Development of the prototype Intercultural-Virtual

Design Team (IC-VDT) modeling and simulation framework for global

project organizations.

Phase 4: Chapter 7: Intellective experiments to examine the impacts of

mixed cultural team cases on team performance

Phase 1: The first phase of the research is to collect data to characterize cultural

differences emerging in joint venture teams of Japanese and American firms, using case

studies and a literature survey. Two points of view adapted from Hofstede (1991) and

Scott (2001), value and practice differences, provide the appropriate framework to

capture the cultural differences that the two different groups bring to a project. Therefore,

I characterize Japanese and American cultures along value-practice dimensions.

Phase 2: The second phase of the research focuses on conducting a cross-cultural

(comparative) study based on Hart’s terminology (1998). I aim to understand and analyze

the impacts of cultural differences on team performance through intellective experiments.

I compare the emergent macro-organization simulation outcomes against predictions of

macro-organization theory in a set of intellective experiments. I use a framework to

understand the effects of cultural values and practices on team performance by

considering three elements: (1) task complexity, (2) organization structure (practice

31

differences), and (3) micro-level behavior (value differences). Four different workflows

are considered for representing different task complexities: pooled, sequential, reciprocal,

and intensive. This research identifies two types of project organization structure that are

linked to Japanese vs. American practice differences, and represent each culture’s

preferred organization structure. It also identifies two patterns of micro-level behavior

that are linked to J-A value differences, and represent preferred behavior of workers from

each culture. The parameters in the current VDT model (using SimVision-R, the

research version of SimVision®) are adjusted to represent both Japanese and American

patterns of micro-level behavior and organization structure. Then intellective

experiments simulate the four possible combinations of micro-level behavior patterns and

organization structures in each project context to predict project work volume, duration,

cost and quality risk outcomes. Emergent simulated results are qualitatively compared to

the two macro-theories: contingency theory (Galbraith, 1974; 1977) and cultural

contingency theory (Hofstede, 1991; Adler, 1997).

Phase 3: The third phase of this research discusses the limitations and the

extensions of the current VDT model to examine the impacts of mixed-cultural teams on

project performance. This phase is important to extend the current VDT model so it can

be used for intercultural studies (Hart, 1998). I develop a prototype model that builds

upon concepts in the current VDT based on my case studies and literature survey. I

describe implementation details of the extended version of the current model, called the

Intercultural-Virtual Design Team (IC-VDT) model.

Phase 4: The fourth phase of this research aims to understand and analyze the

impacts of mixed-cultural teams on team performance through intellective experiments.

In other words, this phase examines and validates whether IC-VDT can be used for

intercultural studies (Hart, 1998). This phase addresses and tests relevant hypotheses

which are proposed as effective leadership styles for mixed cultural teams based on the

literature survey. In particular, the intellective experiment considers four elements: (1)

task complexity, (2) organization styles (Japanese vs. American organization styles), (3)

combinations of micro-level behavior (single cultural team vs. mixed cultural teams), and

(4) matrix strength. The four levels of complexity, organization styles, and

Japanese/American behavior patterns are exactly the same as in Phase 2. The only

32

differences are the combination of team members and the matrix strength. The simulated

results are compared to the hypotheses.

Finally, I discuss the issue of validating IC-VDT, general implications of this

study and its limitations; and I explore future research issues (Chapter 8).

3.4 Research and Validation Process Validation has been a challenging problem in computational organization theory research.

This research adopts an established framework (Thomsen et al, 1999) for validating

simulation models of organizations. The evaluation trajectory proposed by Thomsen et al

specifies a strategy for building up successive validation experiments for new models

(Figure 3.1). Thomsen discussed three major steps of validation: reasoning,

representation, and usefulness.

Reasoning Organization

macro-experience & macro-theory

Simulation macro-behaviors (emergent)

Simulation micro-behaviors

Authenticity Gedanken

Reproducibility

Organization micro-theory & micro-experience

Reasoning & Representation

Reasoning, Representation, & Usefulness

Figure 3.1: Evalua

Note: This figure sal, 1999). A case fof experiments tharepresentation, and

Reasoning:

world behaviors mu

Intellective experiments

Toy Problem

tion Trajectory

hows the evaluation tor the validity of the mt take different approa usefulness of the com

developers of comput

st first validate the m

Generalizability

rajectory proposed bodeling environme

ches. The experimeputational simulatio

ational models that

icro-behavioral assu

Prospective Intervention

nn

a

m

NaturalHistory

y Thomsen (Thomsen et t is built through a series ts test the reasoning,

n framework.

ttempt to emulate real

ptions of their models,

33

either by drawing on empirical micro-social science research findings, or by conducting

their own ethnographies to describe and calibrate micro-behaviors. Drawn or observed

micro-behaviors are calibrated (or “debugged”) using very simple “toy problems” for

which predictions can be simulated manually or with simple spreadsheet calculations.

Second, the predictions of the model for idealized configurations of work processes and

organizations must be externally validated against the predictions of macro theory, via

“intellective experiments.” Specifically, micro-theories relating macro-outcomes to

micro-behavior must match the macro-behaviors observed in the simulation (Thomsen et

al, 1999).

Representation: The second component of the Thomsen’s validation trajectory,

representation, assesses the ability to model and represent a real organization in terms

that make sense to managers (authenticity), that can be replicated by other modelers

(reproducibility), and that apply to multiple projects in different contexts

(generalizability).

Usefulness: Finally, the predications of the model for specific real work must be

externally validated against macro experience via “emulation experiments.” Thomsen

(1999) proposes a “gedanken” approach as the first step to validate the usefulness of

computational models. This step attempts to test the model’s predictions against

managers’ predications. Gedanken experiments build on retrospective evaluation, and

answer “what-if” thought experiments (Thomsen et al, 1999). A gedanken approach is

thus similar to an intellective experiment. The difference is the simulated results of the

idealized or real models are compared to theory and predictions made by managers in the

organization. The final test of usefulness is to test model predications in emulation

experiments against retrospective data from real projects (natural experiments) and

prospectively. When managers develop sufficient faith in the validity of model

predictions, they will begin to make their interventions on model predictions.

VDT has been thoroughly validated in the past for “mono-cultural” projects via

emulation experiments (Levitt et al, 1999; Fridsma and Thomsen, 1998; Nissen and

Levitt, 2004), intellective experiments (Caroll and Burton, 2000; Wong and Burton,

2000), and gedanken experiments (Thomsen, 1998). In order to extend the VDT model

34

toward using it for global projects, this dissertation conducts a couple of intellective

experiments to validate the reasoning assumptions of the IC-VDT model.

Figure 3.2 illustrates the detailed steps for validation of the reasoning assumptions

of IC-VDT. First, this research encodes cultural elements observed during case studies

along value and practice dimensions. The second step is to set up idealized

configurations of work processes and task complexities based on macro-theory. Both

observed elements and idealized project inputs define the inputs of simulations. The third

step is to compare model predictions to idealized project outcomes that are based on

empirical findings in the macro- organization theory and culture literature. The purpose

of this approach is to confirm theoretical consistency. This dissertation completes two

cycles of this validation: the first intellective experiment for single cultural cases and the

second intellective experiment for mixed cultural cases. The two intellective experiments

validate the reasoning of IC-VDT.

Macro-Theory Simulation Observation

Theoretical consistency?

Computational Model Output

Cultural elements Idealized Project Outputs 1) Practices: Organization styles Run Simulation

Macro-TheoryTask control styles VDT / IC-VDT 2) Values:

Define Inputs Idealized Project Inputs Micro-level behaviors

Figure 3.2: Validation of the Reasoning Assumptions of IC-VDT

Note: This figure shows the detailed steps for validation of the reasoning assumptions of IC-VDT. Observations through case studies reveal distinctive cultural elements along value and practice dimensions. Organizational macro-theory can specify both idealized project inputs — idealized configurations of work processes and task complexities — and idealized project outputs — idealized project outcomes of duration and work volume. Both observed cultural elements and idealized project inputs define the inputs of simulations. Finally, model predictions are compared to idealized project outputs, confirming theoretical consistency.

Figure 3.3 illustrates the research processes and the validation steps for the IC-

VDT model. This research starts by creating an additional set of micro-behavior patterns

that represent typical Japanese behavior patterns based on the observations and the

35

literature survey (Phase 1). This assumes that the original micro-behavior patterns

represent American ones, because the current VDT model has been calibrated and

validated against American firms. The second step examines whether the two micro-

behavior patterns are appropriately encoded and represent each culture, using the current

VDT model in an intellective experiment to test reasoning validity (Phase 2). The third

step is to develop the IC-VDT model and to examine ideal mono-cultural teams against

simulated results by the original VDT model (Phase 3). This research assumes that the

current VDT model represents real mono-cultural teams, because of the validation

experiments by many previous researchers for the past 15 years (Thomsen et al, 1999;

Burton and Obel, 1995, etc). The final step is to examine reasoning validity of the IC-

VDT model by conducting the intellective experiment for mixed cultural teams (Phase 4).

(

1) Multiple-Behavior files

tm s

Cross-cultural (comparative) experiments using the current VDT, “Single-cultural” Cases

Limitations and needs (Chapter 4 & 5)

1) Multiple-behavior files

Validity for mono-cultural cases (Chapter 6 & 7)

Define and describe Phase 1 (Chapter 4)

Simulation Phase 2 (Chapter 5)

Theory

2) Organizational practice

Observations

Define and describe Phase 3 (Chapter 6)

Institutional exceptions 2) Organizational practice

National Cultural Index

Figure 3.3: Research Process and Validation Steps

Note: This figure illustrates the research process and the validatVDT model. This research starts to identify multi-behavior pattorganizational practices (Phase 1). These parameters are encodethe current VDT model (Phase 2). In the third phase, this researprototype model, IC-VDT, based on limitations of the current Vobservations (Phase 3). Finally, IC-VDT is validated against bomodel and real “mixed-cultural” team cases (Phase 4).

SimulationResults

Intercultural experiment hrough developing IC-VDT odel, “Mixed-Cultural” Case

Intellective Experiments

ion stepserns (vad and vch develDT modth the cu

SimulationResults

Simulation

- Validity for mixed-culturalcases: Phase 4 (Chapter 7)

flualoer

ValidatedModel

or the IC-es) and idated using ps the l and my rent VDT

36

CHAPTER FOUR: CASE STUDY

This chapter describes the methodology, descriptions, and findings of case studies. The

main objective of these case studies is to understand, characterize, and analyze distinctive

culturally-driven normative systems of Japanese and American teams in IJV projects. In

particular, this chapter characterizes culturally-driven normative systems along value and

practice dimensions. Observed culturally-driven behaviors of individuals are linked to

cultural value dimensions proposed by Hofstede (1991). Empirical findings in culturally-

driven practices are compared to empirical data drawn from the large literature of

Japanese firm-related studies in order to see consistency among industries. Additionally,

this chapter touches on the issue of what are the critical attributes of cultural values and

practices that affect team performance.

4.1 Case Study This research focuses on Japanese and American cultures: First, based on Hofstede’s

research (Hofstede, 1991), there are relatively large differences and gaps between the

two; and second, both Japanese and American construction industries are important

global players (Of ENR’s top 225 international contractors, 42.6% are either Japanese or

American firms). Additionally, these two cultures are also feasible as primary research

subjects, because of my cultural background and the availability and accessibility of

information. I attempt to be impartial in my observations. Accordingly, I often

conducted ethnographic interviews together with an American colleague in order to

observe values and practices from both Japanese and American points of view.

A joint venture (JV) team is the appropriate unit size to observe cultural

differences between Japanese (J) and American (A) organizations. Since each part of the

unit is headquartered in a different country, there is a good possibility that the two sub-

teams are likely to possess their own practices and values, which accumulate through

sharing experiences in their home country’s institutions. When considering that

institutions have not only regulative elements, but also normative and cultural cognitive

elements based on Scott’s definitions (Scott, 2001), the large cultural gap between the

37

two parties implies significant differences in the values and practices of the two

subgroups.

Figure 4.1: A Joint Venture Team

“J” country Regulative systems Normative systems

Cultural-cognitive systems

“A” country Regulative systems Normative systems

Cultural-cognitive systems

Organization A

Joint Venture Team

Team A

Organization J

Team J

Note: This figure illustrates that a joint venture team is assembled by, at least, two or more different subgroups that are headquartered in different countries.

As figure 4.1 shows, this research assumes that sub-teams possess the same

cultural practices and values as their institutional parents. Furthermore, it assumes that

these cultural values represent their respective national cultures. These assumptions are

based on Hofstede’s finding that there are few differences in values between the national

and project-organizational levels (Hofstede, 1991, p.182).

I selected four international joint venture projects between Japanese and American

firms: Semiconductor facility project, Catwalk (C) bridge project, Grasshopper (G)

bridge project, and Sunfish (SF) tunnel project. These cases were chosen, based on data

accessibility and similar scope and scale. Specifically, I fixed the location as around the

San Francisco Bay area to control for the external and exogenous environment

surrounding IJV projects. In the nature of IJV projects, many external factors such as

currency exchange rates, local institutions, and geological conditions make it difficult to

compare projects (e.g., Xiao and Proberbs, 2002). Additionally, the data availability at

available research sites is a substantial constraint to finding Japanese-American joint

venture cases. As table 4.1 summarizes, I conducted six interviews on four different

construction projects using the ethnographic interview approach (Spradley, 1979).

38

Table 4.1: Description of Case Studies

Case study 1 Semiconductor (SC) facility project

Interviewees - “M,” Project Engineer, J firm

- “D,” Senior Vice President, US firm

Project summary Fast-track project (19.5 months)

Design / Build contract

Relationship - Phases 1&2: Conducted by ABC partnership: AAA.

American firm (Architect), BBB. American General

Contractor (Construction), and CCC. Japanese General

Contractor (Structure / MEP Design).

- Phase 3: Conducted by CCC. Japanese General

Contractor with Design/Build contract.

Case Study 2 Catwalk (C) Bridge Project

Interviewees - “Y,” Manager, J firm

- “W,” Project Manager, US firm

Project summary Three span suspension bridge,

3,456 feet long, 400 feet high

Budget $189 Million

Relationship - J fabricator provides the steel deck without erection

- Total about $20 million

- Box type steel deck

- First project to use the box type steel deck in USA

Case Study 3 Grasshopper (G) Bridge Project

Interviewees - “I,” Project Manager, J firm

Project summary Seismic renovation project

Budget $122 Million

Relationship - Joint venture project between J and US firms

- 60% share for the A firm and 40% for the J firm

39

Case Study 4 SunFish (SF) Tunnel Project

Interviewees - “C,” Project Engineer, US firm

Project summary Total 4 km tunnel

Budget About $100 Million

Relationship - Partnership between J and US firms

- J firm is the prime entity contracting with the customer

4.2 Methodology This research uses the ethnographic interview approach, proposed by Spradley (1979),

for data collection. The ethnographic approach is designed for understanding human

culture. The ethnographic interview uses a method of active listening, rather than

proposing a testable hypothesis. The advantage is that interviewees are not influenced,

through the interview questions, by the researcher’s hypotheses or propositions. Taking a

passive rather than assertive role allows the interviewer to observe the unfamiliar cultural

scenes and subtle signals expressed by the interviewee. Thus, the ethnographic interview

approach can be much more effective than the questionnaire approach in terms of

observing subtle information from interviewees (Spradley, 1979) and detecting new

phenomena (Klahr and Simon, 2001).

Appropriate interviews were conducted under the following conditions:

(1) Selection of Interviewees (Informants): Team members who are currently

involved in the projects are the most appropriate informants, because it is easier for them

to recall their experiences. Due to considerations of availability, team members who

were recently involved were also acceptable. Only one person was interviewed at a time

to avoid contamination by other’s opinions. Two interviewees were selected from each

project to compare Japanese and American opinions. For instance, I conducted separate

interviews with one Japanese and one American on the same project, and then compared

their opinions about the same problem.

(2) Duration: I conducted two or three one-hour interviews with each

informant. This provided the necessary time to listen to detailed stories, and was the

maximum duration feasible for the interviewees. Additionally, tow or three interviews

40

were conducted per informant. This procedure enabled me to understand interviewees’

stories clearly and deeply, and to avoid any misinterpretation. For instance, I began by

collecting generic information such as project descriptions, team building, and

organizational configurations. Then, I listened to his or her stories which were related to

cultural differences. At the second interview (typically a week later), I focused on one or

two specific stories which I or the interviewee felt were most relevant. A third interview

was conducted as necessary. These two or three visits allowed me to develop trust

relationships with interviewees and to cover three types of questions- descriptive,

structural, and contrast questions-proposed by Spradley (1979).

(3) Location: Most interviews were conducted in the interviewee’s private

office. Thus, no one else could listen in. This environment enabled interviewees to feel

comfortable and relaxed while speaking. Some interviewees did not care at all, but some

did. This is an aspect of their cultural differences. Noisy places were avoided.

(4) Interviewers (Ethnographers): Two interviewers were used to facilitate

peer review, the sharing of memories, and clarification of language issues. Since the

ethnographic interview approach is very sensitive to nuances of language and nonverbal

signals used by interviewees, the interviewers needed to pick these up precisely and

appropriately. Of the two interviewers used, one spoke Japanese as his native language,

while the other spoke native English, enabling the interview team to understand the

conversation, and also to grasp subtle nonverbal signals, which are deeply related to the

culture.

(5) Interview structure: Spradley suggested three key elements for the

ethnographic interview: the explicit purpose, ethnographic explanations, and

ethnographic questions (Spradley, 1979). The explicit purpose leads to the discovery of

the cultural knowledge of the informants. The interviewers then repeatedly offered

explanations of what they had heard to the informants, to verify that they were learning

and understanding what the informant intended to say8. Ethnographic questions consisted

of three types: descriptive, structural, and contrast questions. By combining the three

8 Informants quite often do not know what are cultural aspects and what are not. The actual conversation hopped around various issues, and sometimes became mixed up, because of our adherence to the unstructured methodology of ethnographic approach. Repeated interviews are recommended in order to clear what kind of cultural differences informants bumped into, how informants feel, and so on,

41

types of questions, the interviewers were able to elicit a variety of useful information

from the informants. The typical sequence of interviews began with descriptive

questions, and then moved on to structural questions, and finally to contrast questions.

(6) Privacy: The collected data is confidential and interviewee names are not

disclosed. Specifically, I made the names of the informants and the projects anonymous,

so that informants could talk freely and avoid risks by telling detailed, true stories.

My case studies were qualitatively characterized and analyzed through a

grounded-theory approach (Eisenhardt, 1989), such as analytic induction (Glaser and

Strauss, 1967) and cross-case pattern search (Yin, 1984). I examined typical behaviors of

team participants and typical practices of cultural teams in IJV projects. Incidents, events,

interpretations, behaviors, and practices were categorized and coded in an iterative

manner, as similarities and differences were noted. Furthermore, observed findings

were compared to the literature (e.g., Hofstede, 1991; Aoki, 1988; Nakane, 1970) to see

differences and similarities.

4.3 Observations The ethnographic interview approach requires the recording of the conversation during

the interviews. These records included rich cultural information, where it was difficult to

discern causes among the various factors such as cultural values, personality, local codes,

and technical issues. The approach employed here is to explain the phenomena using

existing theoretical dimensions, such as those of Hofstede and Trompenaar.

The point of departure for this research, described in the previous chapter, is that

mixed cultural teams bring two types of differences into a project: differences in (1)

values and (2) practices. Observed interview data were broken into the two categories,

and I determined the cultural factors that were key elements to team performance in

construction projects. Cultural dimensions were used to explain cultural phenomena, as

subcategories of value differences.

4.3.1 Value Differences:

Following are qualitative observations of value differences in my case studies.

42

(1) Power Distance Index (PDI): The J firm took the (American) project

manager’s (PM) off-hand comments more seriously than the PM intended (C Bridge

project, Table 4.3:9).

During the coffee break, an American project manager said, “It would be nice

to have the first steel deck section ready to go at 9 am tomorrow morning….”

Next morning, the Japanese firm had it all loaded up and four tugs out in the

water under the bridge and ready to go for 9am….., but they didn't install for

another 3-days. The Japanese engineer said in my interview, “it is a word of

the “Sho-cho (PROJECT MANAGER)!” You know, how can we doubt?

(translated by Horii)”

Japanese members believed that comments or orders originating high in the hierarchy

oblige them to take action, even though the American PM thought that it was just

informal conversation. Additionally, the Japanese manager mentioned “Sho-cho (所長)”

in Japanese. This literally means a project manager, but many Japanese use the word,

“Sho-cho(所長)” with much more respect, since only a senior and professional person

can become “Sho-cho.” This anecdote can be explained by the power distance dimension,

which describes the relationship between bosses and subordinates. Hofstede states that

“less powerful people should be dependent on the more powerful people in the large

power distance countries, while the small power distance countries prefer to have

interdependent relations between less and more powerful people” (Hofstede,1991, p.37).

Thus, the beliefs of the Japanese subordinates are consistent with people from high power

distance countries.

In the SF tunnel project, a Japanese engineer would make sure of the boss’

opinion before making his or her decision (SF Tunnel Project, Table 4.5:6). Hofstede

also categorized decision-making policy using the power distance index. “[S]ubordinates

in the large power distance countries expect to be told what to do,” and “the ideal boss is

a benevolent autocrat or good father in the large power distance countries.” On the other

hand, in the case of small power distance countries, “subordinates expect to be

consulted,” and “the ideal boss is a resourceful democrat” (Hofstede, 1991, p.37).

43

(2) Masculinity Index (MAS): Differences in the MAS in the construction

industry were not observed in the interviews. There are two possible explanations. First,

professional culture in the construction industries of both countries has typically been

very masculine. Alternatively, MAS is not a significant factor on construction projects.

The correct answer cannot be identified in this research. However, it is possible to say

that cultural differences along the MAS dimension do not appear to be a significant factor

in construction projects.

(3) Individualism Index (IDV): In all the case studies, the interviewees

mentioned that Japanese people tend to seek consensus among participants or within

groups. This tendency is one source of conflict between Japanese and American groups.

Specifically, Japanese team decision-making took much longer, since Japanese people

tended to canvas the opinions of all team members. The American engineers felt that the

Japanese decision-making strategy wasted time in getting to a final decision. Hofstede’s

individualism index can explain why Japanese people tend to seek consensus among team

members. In collectivist countries, “harmony should always be maintained and direct

confrontations avoided” (Hofstede, 1991, p.49-78). Based on these observations,

harmony is one of the key points in describing Japanese workplace culture, and can be

seen at many different stages, including meetings and contracts.

In meetings, Japanese people tend to avoid direct confrontation, especially with

the boss or the owner (C bridge project, Table 4.3:8). In the case of the C Bridge project,

since the American consultant represented the owner, the Japanese steel fabricator

avoided arguing during the meeting. Since American meetings tend to be discussions,

silence is taken as agreement. The comments of the American project manager, who

attended every meeting, support this perception. The Japanese kept silent and never

directly confronted the consultant, while the American consultant kept expressing his

opinions and requirements during meetings. The Japanese team brought the agenda,

proposal, or revision to a subsequent meeting, but the consultant would not approve it.

The project team followed this same cycle many times, and the process of steel

fabrication took one year to be approved (C Bridge project). Thus, the Japanese tendency

to avoid direct confrontation in meetings caused misunderstandings for the Americans.

44

The American project manager also mentioned that if the Japanese had said that they

disagreed with the American consultant’s points, things could have been changed much

earlier (C Bridge project).

Japanese firms are relatively careless about signing contracts, especially if they

have a good relationship with the owner. In the Semiconductor project, the Japanese

project manager believed that, because of their good relationship, there would be nothing

disadvantageous to them in the contract, and so signed without careful verification (SC

project, Japanese project manager). The contract did include some disadvantageous

terms for the Japanese contractor, and was one of reasons the project resulted in a

negative profit. The Japanese manager probably thought that their good relationship

extended to the business practice. Sociologists call this approach particularism: treating

one’s friends better than others is natural, ethical, and sound business practice

(Trompenaars, 2004). Hofstede mentioned that particularism and collectivism are

correlated (Hofstede, 1991, p.66-67).

Japanese and American firms take different approaches when the conditions of the

contract have been changed and extra cost has been incurred. The Japanese approach is

to finish the work first, and discuss cost issues later, while the American approach is to

renegotiate the contract before starting work. In the C bridge project, the Japanese

manager said that they could not stop working, even though the conditions of the initial

contract had obviously changed and extra cost was involved, because they did not want to

bother the other workers by canceling their Japanese firm’s jobs only for requiring extra

costs. On the other hand, American firms tend to stop working if they cannot get

agreement about changed conditions or extra cost.

(4) Uncertainty Avoidance Index (UAI): Based on the observations, the UAI

is related to the type of decision and duration of decision making.

One important criterion for Japanese firms is high quality work. In the SF tunnel

and SC projects, the American firm was surprised at what the Japanese team required in

terms of precision and high quality work. The American team sometimes felt that “they

(the Japanese team members) are strict inspectors, rather than engineers.” (SC project,

Table 4.2:9; SF tunnel project, Table 4.5:7) Hofstede mentions precision and quality

45

issues in the UAI index: “In the high UAI and relatively small PDI countries, precision

and punctuality are the most important.” (Hofstede, 1991, p.109-138) Thus, engineers or

project managers of high UAI countries tend to require preciseness and high quality work.

In addition, Hofstede proposed that, “in the high UAI countries, it is important for

a manager to have at hand precise answers to most of the questions that his/her

subordinates may raise about their work” (Hofstede, 1991, p.122). This implies that

project managers of high UAI countries tend to order rework to subordinates; at the very

least, they would not ignore them.

Another tendency is related to the duration of decision-making. For example, in

the SC project, the American manager mentioned that J firm’s decision-making approach

was to make a list of all possibilities, and then choose one by eliminating the others. The

American approach picks a possibility from a small number of choices, and subjects it to

discussion (SC project). Making a list of all possibilities obviously requires a longer

duration.

(5) Long Term vs. Short Term Orientation (LTO): In the SF tunnel project,

Japanese firms contracted with a Japanese machine company to purchase and operate the

shield machine. Based on comments of the American engineer, the Japanese machine

company agreed on a low price with almost no profit for themselves, rather than risk

losing the contract. He explained that there was a certain expectation to win the next

offer from the Japanese contractor if the project succeeded. This story implies that

Japanese firms take a long-term-oriented strategy. On the other hand, this American

engineer felt negative about this strategy, saying that it was not a wise idea because this is

the United States. This suggested that American firms tend to take a short-term-oriented

strategy. Another manifestation of this tendency is that all Japanese general contractors

have ranking and seniority systems, while the American firms use a skill-based reward

system. Ordering relationships by status is one of the characteristics of long term

orientation culture (Hofstede, 1991).

(6) Other value dimensions: Universalism vs. Paticularism Index (UNI)

proposed by Trompenaar is discussed as the other important dimension. As noted

46

previously, Japanese firms tend to enter into contracts based on the relationship, rather

than the written words. Thus, Japanese firms may sign a contract with an owner, even

though the contract includes generic or ambiguous words. Japanese firms tend to think

that ambiguous conditions in a contract can be fixed during the project, for example, at

meetings. In contrast, A firms tend to think that when a contract has been signed, every

condition has been fixed. Therefore, it is reasonable that a contractual document written

by an A firm is long and comprehensive. This difference in business practice relating to

contracts has large potential project risk, because of litigation or mitigation actions. In

order to avoid misunderstanding about changed conditions on a contract, one joint

venture team wrote detailed meeting minutes for each meeting and sent them around to

seek consensus (C bridge project, Table 4.3:8)

4.3.2 Practices Differences

There are the two types of practices: organizational practices and institutional practices.

(1) Organizational practices: Organization practices refer to organization

structures such as the level of centralization, the level of formalization, and the

organizational configuration.

Centralization level: There was no direct expression in the interviews regarding

the most likely degree of centralization for each nation. However, several key signals

were detected from interviews. For instance, Japanese engineers make sure of their

boss’s opinion before issuing decision judgments (SF tunnel project, Table 4.5:4). Also,

J engineers complained that the American project manager did not appropriately give

orders to subordinates. In other words, Japanese engineers expected orders, a sign of

centralization (Semiconductor project, Table 4.3:12). On the other hand, the American

engineer felt that “Japanese engineers are not brave enough to judge by themselves” (SF

tunnel project, Table 4.5:4, 6). According to Hofstede’s power distance index (PDI),

larger power distance countries are most likely to have centralized organizations, while

small power distance countries have decentralized systems (Hofstede, 1991, p.37). Thus,

Japanese firms prefer to have a centralized structure, while American firms prefer a

decentralized structure.

47

Configurations: The Semiconductor and SF tunnel projects, for which the

dominant firm was Japanese, set a multi-level, pyramid-type organizational configuration.

On the other hand, the GC and C Bridge projects, for which the dominant firm was

American, set up a flat and ad-hoc configuration. Hofstede states that “people from a

particular national background will prefer a particular configuration, because it fits their

implicit model, and that otherwise similar organizations in different countries will

resemble one of Mintzberg’s (1980) five archetypal configurations, because of different

cultural preferences” (Hofstede, 1991, p.151). Thus, according to Hofstede’s

classification, Japanese organizations are full bureaucracies with a pyramid model, while

American organizations prefer the divisionalized form. At the project level, since team

members are few in number, the divisionalized form is very close to a flat configuration9.

Thus, Japanese firms are most likely to use a pyramid organizational configuration, while

American firms are more likely to use the flat organizational configuration.

Formalization Level: When two different cultures are involved, meetings tend to

be formalized in order to avoid misunderstandings (C Bridge, Table 4.4:7). For instance,

when both parties realized that they had different approaches to management, meetings,

and negotiations, they would set up mutually agreed-upon rules. Since the construction

project was a one-time event and team members were assembled on a project basis, these

phenomena emphasized formalization. However, when a project team accumulates

experience working together, the formalization level is eased (G Bridge project). Thus,

the formalization level is related to team experience

Hofstede’s work can be used to predict the degree of formalization that would

exist between Japanese and American firms (Hofstede, 1991, p.152). Based on the PDI –

UAI dimensions, high PDI and high UAI countries with full bureaucracies tend to

standardize the work process rather than using mutual adjustment. This means that high

PDI and high UAI countries are most likely to have a highly formalized coordination

system, while low PDI and low UAI countries prefer to standardize outputs. Thus, each

nation has a preference for the formalization level based on their PDI and UAI

dimensions.

9 Hofstede (1991) and Mintzberg (1980) focus on the firm and institutional level, where, in contrast to the project level, larger numbers of people are involved.

48

(2) Task-Control Practices: Task control practices refer to task control

systems such as standardized rules, procedures, and task criterion. The two cultural

groups are most likely to have different practices in rules and goals, which have been

institutionalized in their home countries.

In the case of C bridge project, they discussed the steel fabrication processes such

as welding, the sizes of steel plate, and the specifications of steel panels for a whole year.

The main argument was that the Japanese fabricator tried to use their conventional

processes and specifications which follow the metric system and Japanese standard

systems in terms of efficiency. They pointed out that there is real structural problem to

use Japanese processes and specifications. However, the American design consultant

required them to satisfy the specific number set by the American codes and standards.

The differences in standards are mostly derived from the different systems of units used

in the two countries: metric vs. imperial systems. Additionally, since this project was the

first project to use the steel box type deck for a bridge in the United States, there might

not be appropriate standards in the United States, making their discussion more

complicated. This story implies that the Japanese team stressed process, while the

American team insisted on results and specified use of an American standard. A process-

oriented vs. result-oriented dimension proposed by Hofstede can explain this (Hofstede,

1991, pp189). He argued that in process-oriented cultures people perceive themselves as

avoiding risks, implying a high UAI index. On the other hand, in result-oriented cultures

people perceive themselves as comfortable in challenging goals, implying a high IND

index.

From SF tunnel project and SC facility project, American engineers complained

that Japanese engineers were very picky about quality. This implies that the quality

standard of Japanese contractors is different from that of American contractors. At least,

there is a large gap in expectations regarding quality between the two parties. This can

be linked to the national cultural indexes. In the case of high UAI countries, “precise”

work is respected. Thus, the high UAI countries are most likely to have high quality

standards.

There are many anecdotes to illustrate institutionally-driven practice differences,

such as bamboo scaffolding in Hong Kong, “Tabi (足袋)” shoes in Japan, and so on.

49

These institutionalized practices are customary in their home countries. In a sense,

people may feel that it is taken for granted. Additionally, it is true that it potentially

causes extra cost if they have to change their standard, simply because they have to

purchase new equipment, make hand-made-sized products, or re-educate workers.

4.3.3 Others:

(1) Degree of delegation of power or responsibility: When the American

inspector visited the Japanese steel fabricator to inspect the products, the Japanese

engineer asked the inspector technical questions (C bridge project). However, since the

scope of the American inspector’s work was only to inspect the products, he was not able

to answer any questions. The inspector brought these questions to the United States and

asked the owner, who then asked the design consultant. Finally, answers were sent to the

Japanese steel fabricators a few weeks later. This anecdote implies that the American

delegation’s power is limited because tasks are allocated to many professions. Lincoln

and Kalleberg and Aoki (Lincoln and Kalleberg, 1990; Aoki, 1992) explains this

phenomenon using the “duality” characteristic. They proposed the duality characteristic

of an organization by comparing organizations in two nations, Japan and America.

Japanese firms tend to have a formally centralized structure in their information flow and

a decentralized structure in decision making, called de facto centralization. In the United

States, many tasks and risks are allocated among many groups, such as consultants,

inspectors, sub-contractors, etc, implying that American firms have a formal

decentralized structure in their information flow. However, task allocation decisions are

highly centralized in American firms. In order to conduct an investigation, the American

owner may need to aggregate the information and then re-distribute it again. This

process takes time if a decision needs to cross multiple professions. In contrast, since

Japanese engineers tend to be assigned to multiple tasks, the process is relatively short.

(2) Availability of human resources: It is relatively difficult for foreign

companies to maintain skillful foremen and superintendents and to have skillful

employees who are well versed in American business practices, since procurement of

projects is uncertain (SC project, Table 5.2:7, 8). In the case of the C bridge project (C

50

bridge project, Table 5.3:1), the project manager pointed out problems resulting from the

lack of a feedback system in the Japanese firm.

(3) Team Experience: Cultural conflicts and misunderstandings are sensitive

to team experience. In the G bridge project, the Japanese firm had had experience

working with the same American firm. There is apparently no significant problem with

their performance. Additionally, the Japanese project manager had had experience in the

United States, including education at an American university. He mentioned that he had

tried to omit any Japanese business styles in the project, which was one of the key factors

that led to project success (G Bridge project, Table 5.4:1,2).

Table 4.2: Summary of Interviews - Case Study 1: Semiconductor Project

Differences Japanese (J) Firm American (A) Firm

1 Different

approaches to

decision

making

Consensus among participants

was the first priority, after

which the J team could make

a decision.

A team members tended to make

decisions by themselves. A team

members preferred “cowboy

style” rather than “group”

decision-making.

2 Different

approaches to

making offers

when the

conditions of

the contract

have been

changed and

extra costs

have been

incurred.

The J firm tried to finish the

job first, and then filed the

claim, because the J firm

worried about affecting

interdependent work through

stoppage. Thus, from the

point of view of the A firm,

claims made by J firms

tended to be late.

The A firm stopped working if

there was a conflict, unless the

contract was modified based on

“changed conditions.”

51

3 Different

expectations

about

compensation

The J firm expected that

“Seii” (faithful or sincere)

work would be evaluated (and

compensated) by the owner.

The J firm first demonstrated

their sincere effort, and then

asked for changes in the

conditions.

To the A firm, the contract was

final. The A firm believed that

an offer to change conditions

should be made before work

begins or the next task is begun.

No compensation should be

given unless there is description

in the contract.

4 The meaning

of meetings

To the J firm, a meeting was a

place to make decisions or

reach consensus among

participants, rather than a

discussion forum.

To the A firm, a meeting was a

place for discussion, rather than

for reaching consensus.

5 Language

barriers

J teams could not

communicate well in

meetings because of the

language barrier. (Even a

fluent speaker spoke less

frequently than A meeting

participants.)

To A, the silence of J

participants was a signal of

agreement.

6 Different roles

in the

organizational

structure

The project manager and

project engineers were

responsible both for

paperwork and for managing

the field work. Many J

engineers helped to manage in

the field, if the

superintendents were not

skillful.

The project managers and

project engineers tended to focus

on paperwork, especially

contract issues. The

superintendent took care of the

construction site. Overlapping

of work responsibilities was less

likely.

52

7 Difficulty of

hiring and

retaining

skillful people

Since it is relatively difficult

to constantly acquire new

projects in the United States,

it was hard to keep skillful A

superintendents and J

engineers who have had

experience in the United

States

The A firm took advantage of

local firms. The A firm could

easily hire skillful

superintendents who have had

vast experience in the local area.

8 Feedback

problems

caused by

inconsistent

project

acquisition

The J firm failed to get

feedback from past projects,

due to inconsistent project

acquisition in the United

States

Feedback system was working

well in A firms

9 Quality level

required

To J firms, high quality work

and products were given first

priority, even if the quality

level exceeded that specified

in the contract.

To A firms, quality was defined

by the contract. The goal of the

A firm was to satisfy the quality

level defined in the contract.

10 Relationship

between the

general

contractor

(GC) and sub-

contractors

The GC had more power than

the Sub. The relationship

between the two was similar

to a boss-employees or father-

sons relationship. Thus, it is

rare that a Sub sues the GC.

Subs were treated as equal to the

GC. Subs commonly sue the GC,

and vice versa.

11 Relationship

between GC

and owner

The owner was treated like a

king. Their orders were taken

absolutely, even if they were

unreasonable and out of the

scope of the contract

The relationship was relatively

equal.

53

12 Management

style

J firms used a relatively

centralized and formalized

system, since team members

made sure of the boss’

opinions each time.

However, one problem arose

because the project manager

lacked experience in the

United States

A engineers worked

independently. The A engineer

subordinate to the J project

manager preferred to work

freely. The preference was for a

relatively decentralized system.

13 Decision

making process

The decision making process

sought to eliminate less

feasible options, and to decide

upon the single best option.

The A team came up with three

or four ideas and quickly

decided upon one option. Then

they discussed and modified that

option.

Table 4.3: Summary of Interviews - Case Study 2: Catwalk Bridge Project

Problem Japanese (J) Firm American (A) Firm

1 Feedback

problems

caused by

inconsistent

project

acquisition

Although the J firm had had

over 10 projects in the United

States, there was no feedback

and training program from the

past projects.

This was the first project to

contract with the J firm as a

subcontractor. There was no

data about cultural issues.

2 Technology

level

The J firm possessed the

ability to produce high quality

products. This project was the

first bridge project to use the

box type steel deck in the

United States.

The technical expertise of the J

firm was the main reason to

contract with the J fabricator

instead of another firm

54

3 Longer

approval time

from the A

consultant (1

yr)

To the J firm, technical issues

and the production process

was the responsibility of the

fabricator. The fabricator

proposed a higher quality and

cheaper method to the J firm.

However, the A consultant did

not understand, simply

because of documentation

problems and unfamiliarity

with the techniques. This was

a linguistic problem rather

than a technical or quality

issue.

All concerns raised by the

consultant were related to

process documentation rather

than the products. For instance,

the consultant pointed out

grammatical errors, rather than

the content. In addition, the

consultant did not know the new

technology and method, and

required multiple explanations

4 Language

barriers

The J team kept silent in the

meetings. The J team agreed,

more or less, with the

grammatical errors pointed

out by the A consultant.

To the A consultant, silence was

a signal of agreement. The A

consultant spent much time

correcting the written English in

the documents.

5 Role of

inspectors

Generally, since the inspector

was the same person as the

project engineer (project

manager), he had the power to

decide technical issues at the

factory.

Inspectors coming to Japan did

not have any authority to make

decisions at the factory when the

J firm asked questions regarding

technical issues. The purpose of

the inspectors was simply to

inspect products. Thus, the

delegated power to the inspector

was very narrow and limited.

55

5 Different

approaches to

change order

requests

When conditions changed, the

J firm tried first to finish the

job, and then asked for extra

compensation for the

changing conditions. Thus,

the J firm’s approach was to

complete the work first, then

think about cost.

The A firm tended to stop work

relatively easily if there was a

conflict about the changing

condition of the contract. Thus,

the A approach was to reach

contract agreement first and to

finish work afterwards.

6 Beliefs and

values “Seii (誠意)” (a faithful or

sincere work attitude) was

quite important to achieve a

good relationship among the

parties. Sometimes faithful

work was sufficient reason for

compensation from the

owner.

Faithful work could achieve a

good relationship, but was not

going to be a basis for

compensation. To the A firm,

the contract conditions stood.

7 Formalization

level of

meetings

Since the J firm wanted to

avoid misunderstanding, they

made meeting minutes and

distributed copies to the

participants to make sure of

the meeting contents.

The A firm also tried to avoid

misunderstanding and made an A

version of the meeting minutes.

8 Meaning of

meetings

Meetings were a place for

reaching consensus or making

decisions. The J firm tried to

avoid direct confrontation

Meetings were an opportunity

for discussion rather than for

reaching consensus about change

orders or the conditions of the

contract.

9 Power distance The J firm took a comment

from the A project manager to

call tug boats as a serious

order.

The A project manager believed

it to be a casual conversation,

not a genuine order.

56

10 Differences in

meeting styles

and contract

definitions

Meetings were the place to

make the ambiguous contract

clear. To the J firm, contracts

use many general terms.

Traditionally, J contracts run

only a few pages long.

When both parties agreed with

the contract, every condition

stays fixed even if there are

generic or ambiguous words.

Thus, if one firm signed the

contract with an ambiguous

condition, it is that firm’s

responsibility.

11 Meeting

minutes

For the reason listed above,

meeting minutes were

important in Japan. The

minutes specified who said

what and the final agreements

on issues.

Meeting minutes were

important. However, they were

not enforceable without

agreement or signature, because

the purpose of meetings is the

discussion of issues.

Table 4.4: Summary of Interviews - Case Study 3: Grasshopper Bridge Project

Problem /

Tendencies

Japanese (J) Firm American (A) Firm

1 Successful

factors

The J manager tried to use the

A management method, such

as the “cowboy style”.

The A firm had a flat

organization, less formal

meetings, and a good

relationship between the J and A

firms.

2 Experience The J manager had extensive

experience and an educational

background in the United

States.

Both firms have had experience

working together

3 Team size The J team consisted of two

individuals, and thus the

magnitude of the cultural

impact may have been

The A team consisted of eight

individuals.

57

lessened.

4 Uncertainty

avoidance

The J manager tended to be

concerned about information

disclosure.

The A manager tended to share

information openly.

Table 4.5: Summary of Interviews - Case Study 4: SunFish Tunnel Project

Problem /

Tendencies

Japanese (J) Firm American (A) Firm

1 Problems

regarding the

contract

Based on the J business

custom regarding joint

ventures, the dominant firm

had a “sponsor merit” share

of the profit.

The A firm believed that both

firms had reached agreement on

profit sharing.

2 Litigation

problem

The J firm did not expect that

the A firm would sue the J

firm, a partner company.

The A firm believed that

litigation was only one way to

solve this problem and

misunderstanding, because it

was a breach of contract.

3 Differences in

the relationship

between the

general

contractor

(GC) and

subcontractors

(Subs)

In the construction business in

Japan, the relationship

between a GC and Subs was

similar to the relationship

between father and sons or

daughters.

In the construction business in

the United States, the

relationship between a GC and

Subs is equal.

4 Time for

decision

making

The J engineers asked their

boss’ judgment before asking

their subordinates to carry out

any orders.

The A engineers felt that the J

engineers could not make certain

decisions. The A engineers felt

that the J process required

58

unnecessary decision making

time and was inappropriate for

the construction industry.

5 Response to the

owner’s order

The J firm demonstrated an

effort to satisfy the owner’s

orders as much as possible.

The A engineer rejected the

owner’s order because it looked

like it was causing low

productivity and schedule

overruns.

6 Delegation of

power

The J group tended to use

consensus among team

members, rather than delegate

power.

The A engineers felt that the J

engineers were not empowered

to make decisions.

7 Preciseness and

quality issues

The J firm wanted to provide

high quality work to the

owner. Issues pointed out by

the J engineers were mostly

related to the quality.

The A engineers felt that the J

engineers resembled

“inspectors” more than

“engineers.”

4.4 Conclusion and Discussion This research began by defining the characteristics of Japanese and American teams

through ethnographic interviews, observations, and the literature survey. To summarize,

many anecdotes collected through ethnographies demonstrate that there are, by and large,

cultural differences between Japanese and American teams in IJV projects. I observed

mainly two types of differences: values and practices.

My ethnographies show that Japanese and American participants have their own

distinctly different patterns of micro-level (individual) behaviors (Table 4.6), in particular,

in decision-making and communication. Japanese workers, for instance, tend to seek

consensus before making decisions, while Americans prefer to decide independently.

Each individual behavior pattern is explained by using Hofstede’s cultural value

dimensions (Hofstede, 1991). For instance, the Japanese culture that has the low score

59

(46 out of 100) of the individualism value index can give a reason of why Japanese

participants insist on group consensus for their decision making. On the other hand, the

American culture gets the highest score (91 out of 100) among countries in the

individualism index, representing individual-based decision-making and communication

behaviors. Therefore, cultural value dimensions can provide not only interpretations of

participants’ behavior patterns, but also potentially the degree of value gaps between

countries. However, there is no difference observed in the masculinity index between the

two cultures, although raw score gap between the two countries are large (Table 2.1).

This implies that some cultural value indexes may not be critical in a certain industry.

Based on above findings, this dissertation redefine “values” as desirable criteria or

standards for evaluating behaviors that people show in making task-related and

communication-related decisions

The second finding is related to differences in organization control practices. My

observations reveal that Japanese and American teams in IJV projects have brought their

own practice styles into the project. For instance, Japanese project teams tend to have

multiple levels of hierarchy and to be more centralized, while American firms usually

adopt a flat organizational hierarchy and decentralized authority. These observed

tendencies are consistent with existing literature (e.g., Nakane, 1970; Ouchi, 81: Sullivan

and Nonaka, ’86). Nakane (1970), for instance, emphasized that the hierarchical

structure is the fundamental structure of Japanese organization systems as appeared in

ranking systems and seniority systems, because of the strong “Senpai-Kohai (Senior-

young)” relationship (Nakane, 1970). Aoki (1988) and Lincoln and Kalleberg (1991)

found that Japanese firms tend to have high levels of formal centralization, confirming

my observations. In addition, these practice styles can be explained by Hofstede’s value

dimensions. PDI scores, for instance, support the level of centralization and the

organizational configuration in both Japanese and American teams.

The third finding is related to task control practices such as rules, criteria, and

standards. Japanese and American firms have differing task control practices. Japanese

construction firms, for instance, try to control tasks by processes, while American firms

emphasize results. The differing task-control practices require team members to discuss

which standards the IJV project team will select. This type of exception must be

60

distinguished from technical and project exceptions, since it is neither technically related

matters nor integration issues on products. I call these institutional exceptions. A set of

rules and standards of each firm has been shaped and fostered for years under the various

institutional environments such as home country’s regulations, professional norms,

corporate and national cultural values and norms. In other words, these standards and

rules have been institutionalized for years or have been adapted to their home country’s

institutions. Interestingly, differing task control styles can also be linked to Hosftede’s

value dimensions. He argued that high UAI countries tend to avoid uncertainties by

controlling tasks by processes (process-oriented vs. result-oriented cultures). Therefore,

the large gaps in the two countries’ scores have the potential to cause institutional

exceptions.

By combining the organizational control and the task control practices, this

research extends the meaning of “practices” to include cultural norms for adopting or

using specific organization designs to manage organizations and tasks.

Table 4.6: Summary of Findings

Values Culture J (Japanese) Culture A (American)

Decision

Making

Group based decision making Individual based decision making

Communication Group-based

communication

Individually-based

communication

Practices Culture J (Japanese) Culture A (American)

Centralization Centralized authority Decentralized authority

Formalization High level of formalization Medium level of formalization

Organizational

hierarchy

Multiple levels of Hierarchy Flat level of hierarchy

Task control

styles

Control by processes Control by results

61

To summarize, each cultural team composing an international joint venture team

will have its own set of typical values and typical practices, named culturally-driven

normative systems, which are the typical coordination mechanisms of a Japanese team vs.

an American team (Table 4.6). Misunderstandings and miscommunications might be

results of not only lack of awareness and knowledge of the other party’s normative

system, but also struggles to find a legitimate compromise between the two systems.

Additionally, several differences were not categorized as practices or values

differences. These are related mostly to circumstances surrounding the project team,

such as the availability of human resources and team experience. These environmental

factors usually make a situation worse, since a lack of individual and team experience

implies increasing the information processing demand for the team.

62

CHAPTER FIVE: INTELLECTIVE EXPERIMENTS FOR

SINGLE-CULTURAL TEAMS

The previous chapter confirmed that the Japanese and American teams have their own

sets of culturally-driven normative systems in IJV projects. As the second phase of the

research, this chapter focuses on conducting a cross-cultural (comparative) study based

on Hart’s terminologies (1998). The main purposes of this chapter are (1) to encode

culturally-driven normative systems using the information processing abstraction in the

Virtual Design Team (VDT) model, and (2) to understand and analyze the impacts of

culturally-driven normative systems on project performance through an intellective

experiment. The intellective experiment compares emergent simulated results to the

predictions of macro-organization theory. Culturally-driven normative systems are

broken into two main factors: cultural values and cultural practices. My first intellective

experiment attempts to look into and compare the impacts of each cultural value and

practice, before examining mixed cultural cases in Chapter 6 and 7. I preliminarily focus

on behavioral patterns of project participants as cultural values and organization

structures as cultural practices (Chapter 4.3.2) due to the limitations of the current VDT

model. Emergent simulated results are qualitatively compared to the two macro-theories:

contingency theory (Galbraith, 1974; 1977) and cultural contingency theory (Hofstede,

1991; Adler, 1997). Thus, there are four sub objectives:

- Understand effects of changes in the organization structures (practices)

- Understand effects of changes in the behavioral patterns (values)

- Understand relationships between the organization styles and behavioral

patterns

- Analyze theoretical consistency of emergent simulated results

- Assess and validate qualitatively encoded cultural values and practices

63

5.1 Simulation Models as a Methodology How can we determine and observe the effects of these cultural factors? In a real project,

there are so many factors involved that it is really hard to determine pure effects of

cultural factors on team performance. On the other hand, computer simulation is growing

in popularity as a methodological approach for organizational researchers (Dooley, 2002).

In other words, computer simulation provides a “virtual laboratory” where we can

address a question about organization science. Computational laboratories permit us

greater experimental variety to complement other approaches used in organization

science (Burton, 2003). Specifically, many researchers have argued that simulation

models allow researchers to examine a series of “what-if” questions (e.g., Dooley, 2002;

Burton, 2003; Carley, 1995; 1996) in addressing appropriate questions related to the

purposes of the model. Based on Burton (Burton, 2003), the real world is a “big”

laboratory where we essentially have a single run observation — a stream of data over

time, which we cannot re-run. Computational models take and represent some part of the

“big” laboratory, but allow us to re-run experiments again and again with subtle

differences in inputs and thereby to gain insights about cause-and-effect relations10 by

simulation experiments.

Can we use the VDT model as a laboratory? Are the questions this research

raised relevant to the objectives of the VDT model? Burton & Obel “(1995) discussed

that if a model satisfies three validation steps — reality, content, and structure — the

model can be used to examine hypotheses, operating as a “theorem prover,” so long as

tested hypotheses are appropriately relevant to the objectives of the model.

The reality validation for VDT has been proven through previous research (Levitt

et al, 1994; Levitt et al, 1999). Since the VDT model already closely represents real

projects; we assert that it can be used to confirm or reject theoretical predictions by

modeling and simulating sets of idealized organizations.

For construct validity, existing parameters and variables must adequately

represent theory. Since this research links cultural factors and VDT parameters, based on

10 There are other advantages discussed by Lin and Carley (2003, pp. 30-31) such as no damage to the existing environment, no bias by looking at successful projects, comparability, and identifying dominant factors affecting performance.

64

observations and literature review, this step is satisfied if this framework is used for

qualitative analysis.

The final step, criterion-related validity, demonstrates whether the VDT model

matches the purposes of the theory. The main purpose of the VDT model is to predict

team performance in project organization design based on information processing theory.

This research also intends to understand effects of changes of project-organization design.

Additionally, much of the empirical data used in calibrating the VDT model comes from

facility engineering projects, which is the same project type on which this research

focuses.

Additionally, Hofstede’s theory of “the preferred coordination mechanism”

straddles two fields, organization theory (Mintzberg, 1980) and sociology (Hofstede,

1991). Similarly, two sets of organization designs, organization structures and behavior

patterns, are observed and manipulated in the proof-of-concept model based on the two

fields, organization theory and sociology. Therefore, the VDT model provides a useful

laboratory to observe the qualitative predictions of the theories.

5.2 Modeling Can the information processing abstraction in VDT appropriately represent and capture

cultural factors? How can we appropriately encode value-and-practice related parameters

based on observations?

This section begins by discussing the VDT simulation system. Figure 5.1

illustrates input and output variables for a VDT simulation (Jin and Levitt, 1996). VDT

differentiates three types of variables: organization descriptions, project contextual

descriptions and project performance. Organizational descriptions define organization

design — e.g., level of centralization, level of formalization, and use of communication

tools. These parameters are controllable by project managers. For a given project

context description, a project manager may change organization designs to see their

effects on project performance. Therefore, organizational descriptions can be defined as

decision variables or independent variables in this experiment. Project contextual

descriptions define a given project — e.g., activity, interdependency among activities,

and number of actors. Project contextual descriptions are generally given, thus seen as

65

control variables. Dependent variables are outputs of each simulation run and change as

a function of the independent and control variable settings. The VDT simulation outputs

include project duration, cost, and quality risk.

Figure 5.1: Inputs and Outputs of VDT Simulations: adapted from Jin and Levitt (1996)

My ethnographies found that Japanese and American firms are most likely to

bring their own sets of culturally-driven normative systems into an IJV project (Chapter

4). Culturally-driven normative systems are broken into the two main factors: cultural

values and cultural practices. Cultural values are linked to descriptions of project team

members, while cultural practices are linked to organization structure through VDT input

variables (Figure 5.1).

Observed differences in cultural practices through case studies indicate that

Japanese and American firms have their own preferred organization structures (Chapter

4). An organization structure represents the coordination mechanism to control or

constraint individuals’ actions in making decisions and/or communicating among

Organizational descriptions:

Pro ect contextual descriptions:j

Organization structure (Each cultural group has its own organization style) Communication tools

Descriptions of project team members (Each cultural team member shows its own individual behavior pattern) Descriptions of activities and their dependencies

INPUTS OUTPUTS

Project Performance

(Dependent variables)

VDT Project duration

Simulation Project cost Project quality

Note: This figure illustrates input and output variables for a VDT simulation (Jin and Levitt, 1996). Input variables are broken into the two categories: organizational variables and project contextual variables. Organization descriptions define organization structure and communication tools (independent variables). Project contextual descriptions define a given project such as team members, project activities, and interdependencies among activities (control variables). Outputs of a simulation include project duration, cost, and quality risk.

66

individuals (Baligh and Damon, 1980; Baligh and Burton, 1981; Malone, 1987). For

instance, in a highly centralized structure, most decisions are made at the top of the

control structure by project managers. Thus, when an engineer detects an exception, the

engineer is most likely to report the exception to his/her sub-team leader, and the sub-

team leader passes the exception to a project manager for a decision. In other words,

centralization policy and centralized control structure enforce constraints on actor’s

decision-making actions in processing information. Formalization policy and formalized

communication structure affect communication actions among individuals (Jin and Levitt,

1996). For instance, while a communication structure defines who can talk to whom, the

level of formalization of an organization defines how frequently they will send written

communications to each other. Therefore, differing cultural practices observed through

case studies suggest having two sets of organization structure styles that represent

Japanese vs. American leadership styles.

Observed differences in cultural values indicate that Japanese and American

participants show different behavior patterns in decision-making and communication

(Chapter 4). The VDT model sees individual actors as the minimum unit of information

processing. Actors are assigned a certain type of behavior pattern, called micro-level

behavior of actors, which refers to the actor’s decisions about how to process and

communicate information. In the VDT model, actors have one set of behavior patterns

that depends on the actor’s role, one of project manager (PM), sub-team leader (SL), or

sub-team member (ST). Differing cultural values observed through case studies suggest

having two sets of behavior patterns that represent Japanese and American project

participants.

The above discussions indicate that observed cultural factors can be represented

and captured using the information processing abstraction in VDT. The next step is to

design an intellective experiment that attempts to understand and analyze the effects of

culturally-driven normative systems on project performance. Specifically, the first

intellective experiment looks into the impacts of changes in the values and practices. The

intellective experiments follow the three steps as discussed in Chapter 3 (Figure 3.2).

First, this research encodes cultural practices and cultural values observed through

case studies. Cultural practices and cultural values are linked to organization structures

67

and micro-level behaviors respectively. In particular, this research sets up two types of

organization structures: Japanese organization style and American organization style.

Similarly, this creates two sets of micro-level behavior patterns: Japanese behavior

pattern and American behavior pattern.

The second step is to set up idealized project configurations of work processes

and task complexities based on macro-theory. Specifically, this research addresses four

basic types of idealized project configurations (Thompson, 1967; Bells and Kozlowski,

2002): a pooled workflow, a sequential workflow, a reciprocal workflow, and an

intensive workflow. Additionally, this research examines different levels of team

experience as representing a real IJV project team situation. Therefore, this first

intellective experiment has four input variables: (1) organization structure, (2) micro-

level behavior of actors, (3) idealized project description, and (4) team situation.

The third step is to compare model predictions to idealized project outcomes that

are based on empirical findings in the macro-organization theory and culture literature. In

other words, the organizational and micro behavior parameters are incorporated and

tested as a proof-of-concept experiment. Given the limitations of the current VDT model,

this intellective experiment examines only mono-cultural team cases and does not

incorporate task control styles. These elements are considered in Chapter 6 and 7.

5.2.1 Organization Structure

Each national culture has its preferred coordination mechanism (Hofstede, 1991). This

implies that Japanese and American firms have their own preferred organization structure,

independent of the task complexity and team circumstances. Thus, there are two types of

organization structure: Japanese and American styles. Each style is composed of four

elements: centralization, formalization, matrix type, and organizational configuration.

Table 5.1: Leadership Styles as Organization Structure

Leadership Style Type J Type A

Centralization High Low

Formalization High Med

Organizational Configuration Multiple layers of hierarchy A flat level of hierarchy

68

Based on the observation and the literature review, the Japanese style can be

categorized as a tight system, because they generally have high centralization, high

formalization and multiple levels of hierarchy. The American style is a relatively loose

system, identified by medium centralization, formalization, and matrix level, and a flat

configuration.

5.2.2 Individual Behavior Pattern

Based on the literature survey and my ethnographies, three major cultural indexes and six

main behavior parameters are selected. Three cultural indexes – power distance index

(PDI), uncertainty avoidance index (UAI), and individualism index (IND)11 (Hofstede,

1991) – are considered important cultural values contributing to micro-level behavior

patterns. Behavior patterns are determined by six relevant behavior parameters: 1)

decision making policy, 2) type of decision (rework, correct, or ignore), 3) tolerance in

waiting for decisions, 4) attendance to communication, 5) response volume of

communication, and 6) demand volume for communication. Parameters 1-3 relate to

decision-making behaviors. Parameters 4-6 relate to communication behaviors. The

relationship between behavior types and the national cultural index is shown below

(Table 5.2).

Table 5.2: National Cultural Index – Behavior Matrix

PDI UAI IND

Decision making policy + N -

Type of decision (Rework, Correct, Ignore) + - N

Decision

Making

Tolerance of waiting time for decision + N -

Attendance to communication N + -

Response volume of communication N + -

Communi-

cation

Demand volume of communication N + N

Note: Table 5.2 shows the matrix of national cultural value indices vs. micro-level behavior parameters. (”+” represents positive correlation. “-“ represents negative correlation. “N” indicates no correlation)

11 MAS index was not observed from case studies. A possible explanation is that both workers in the construction industry have relatively the same level of MAS. LOT index has relatively less impact on project because of the short term focus in all cultures on the events that drive project decision making.

69

(1) Decision-making policy: This matrix determines which actor should make

the decision for an exception, based on the project's centralization level, power distance

index, and individualism index. The PDI refers to boss-and-subordinate relations.

According to Hofstede, “subordinates in large power distance countries expect to be told

what to do,” and “the ideal boss is a benevolent autocrat or good father in the large power

distance countries.” On the other hand, in the case of small power distance countries,

“subordinates expect be consulted,” and “the ideal boss is a resourceful democrat”

(Hofstede, 1991, p.37). Thus, power distance increases the tendency toward formal

centralization. Formal centralization refers to formal authority representing the

information flow from subordinates to middle and senior managers. As an example,

Lincoln and Kalleberg (1990) observed that all documents in Japanese firms wend their

way up the hierarchy, which is correlated to high power distance.

Preposition 1: The higher the power distance index (PDI), the greater the

centralization of decision-making behavior.

On the other hand, some researchers (e.g., Lincoln and Kalleberg, 1990; Aoki,

1994, 1988) characterized the Japanese decision-making system as bottom-up rather than

top-down, and thus mismatched to the power distance and formal authority features. The

“ringi” system in Japan is famous as a way to make decisions by seeking consensus

among group members. In other words, the actual decision making is decentralized

among the group members. Lincoln and Kalleberg (1990) called this “de facto

centralization,” as distinguished from “formal authority.” “De facto centralization” is

linked to the individualism-collectivism dimension. The individualism index proposed

by Hofstede can explain why Japanese people tend to seek consensus among team

members. In collectivist countries, “harmony should always be maintained and direct

confrontations avoided” (Hofstede, 1991, p.49-78). Based on these observations

(Hofstede, 1991; Ouchi, 1980, etc), harmony and trust among group members are key

aspects of Japanese workplace culture, and can be seen in many different ways, including

meetings and contracts. Thus, the lower individualism, high collectivism countries tend

to have low de facto centralization of decision-making.

70

Preposition 2: The lower the individualism index (IND), the less de facto

centralization in decision making.

Moreover, the total centralization level needs to take into consideration the

balance between formal and de facto centralization. The initial effects of formal and de

facto centralization are equally distributed. Therefore, the decision-making for an

exception behavior (decision maker policy12) is calculated as follows:

DecisionMakerPolicy (X) = 0.5 x PDI_FormalCentralizationFactor + 0.5 x

IND_DeFactoCentralizationFactor

I begin by changing parameters of micro-level behavior files qualitatively in the

VDT model. I assume that the default version of the VDT model represents American

behavior patterns. Therefore, I can set up parameters assuming that the American scores

of the cultural indexes can always be the default version of the micro-level behaviors in

the VDT. For instance, the American PDI score is 40, in the range of between 25 and 50,

representing the default value of the VDT and providing a benchmark probability for

adjusted probabilities depending on PDI score. Then, I can adjust parameters of the

Japanese micro-level behaviors by measuring the relative score gaps from the American

score of cultural indexes. For instance, the large score gap in PDI index can increase or

decrease a probability that a higher level participant makes a decision. The range of

adjustment would be from plus 20% to minus 10%. This research begins by calibrating

encoded parameters with 20% differences as the largest score gaps relative to the

American value scores for making qualitative differences in individual behaviors.

Although quantitative research remains as future research work, it is interesting to see the

effect of qualitative differences on project performance.

12 This term, decision maker policy, is used in the current VDT model to determine decision making behavior of actors.

71

Table 5.3: Adjustment Factor for Decision Making Policy

Adjustment factors for

Decision Making Policy

X > 75 75 >X > 50

(Pattern J)

50 > X > 25

(Pattern A)

25 > X

PM +20% +10% 0% -10%

SL 0% 0% 0% 0%

ST -20% -10% 0% +10%

PD

I >7

6

75>P

DI>

51

50>P

DI>

26

25<P

DI ST

SLPM

00.1

0.20.3

0.4

0.5

0.6

0.7

0.8

ST

SL

PM

Figure 5.2: Adjusted Probabilities for Decision Making Policy (High Centralization)

Note: Table 5.3 indicates the adjustment factors for decision making policy. Figure 5.2 show the adjusted probabilities for who makes a decision about an exception in the case of high centralization. As can be seen, the higher the PDI score becomes, the higher the probability that higher level participants (PMs or SLs) resolve the exception. Since the numbers in the table represent probabilities, the sum of probabilities in PM, SL, and ST should be equal to 1. The American PDI score is 40, in the range of between 25 and 50, representing the default value of the VDT and providing a benchmark probability for adjusted probabilities depending on PDI score.

(2) Type of decision (Rework, Correct, or Ignore): This matrix is used by a

decision-maker to determine how an exception should be dealt with, based on the

project's centralization policy, power distance index (PDI), and uncertainty avoidance

index (UAI). In the high PDI countries, a manager is more likely to provide the solution

rather than using a consulting approach (Hofstede, 1991), implying an increase in

"Correct" decisions. In the high UAI countries, there is a high awareness of "Ignore"

decisions. Hofstede mentions precision and quality issues in the UAI index: “In the high

UAI and relatively small PDI countries, precision and punctuality are the most

72

important” (Hofstede, 1991, p.109-138). Engineers or project managers of high UAI

countries tend to require preciseness and high quality work. Thus, high UAI countries

show fewer “Ignore” decisions than low UAI countries.

Preposition 3: High UAI countries show fewer “Ignore” decisions than low UAI

countries.

Preposition 4: High PDI countries show more “Correct” decisions than low PDI

countries.

Table 5.4 shows the adjustment of decision types for an exception. The actual

probability of decision type is calculated as: Adjustment factor x Defaulted Probability.

The adjustment factors are determined by UAI and PDI scores of each country. The UAI

score influences probabilities of both “Rework” and “Ignore” decisions, while the PDI

score determines probability of “Correct” decisions. Figure 5.3 exemplifies how much

the probabilities are deviated from the default probabilities. The default version of VDT

shows 0.65 for “Rework,” 0.3 for “Correct,” and 0.05 for “Ignore” decisions, in the case

of centralized authority. These probabilities range between-10% and +20% depending on

UAI and PDI scores’ relative differences to American scores.

Table 5.4: Adjustment Factor for Decision Types

Probability UAI > 75

(Pattern J)

75 >UAI > 50 50 > UAI > 25

(Pattern A)

25 > UAI

“Rework” decisions

(PM, SL, ST)

+20% +10% 0% -10%

“Ignore” decisions

(PM, SL, ST)

-20% -10% 0% +10%

Probability PDI> 75 75 >PDI > 50

(Pattern J)

50 > PDI > 25

(Pattern A)

25 > PDI

“Correct” decisions

(PM, SL, ST)

+20% +10% 0% -10%

73

Probabilities of Decision Types by PM

0

0.2

0.4

0.6

0.8

1

Rework Correct Ignore

0.3

0.65

Ranges of adjusted probabilities

Defaulted probabilities

0.05

Figure 5.3: Adjusted Probabilities for Decision Types

Note: Table 5.4 shows the adjustment of decision types for an exception. The actual probability of decision type is calculated as: Adjustment factor x Defaulted Probability. The adjustment factors are determined by UAI and PDI scores of each country. The UAI score influences probabilities of both “Rework” and “Ignore” decisions, while the PDI score determines probability of “Correct” decisions. Figure 5.3 exemplifies how much the probabilities are deviated from the default probabilities. The default version of VDT shows 0.65 for “Rework,” 0.3 for “Correct,” and 0.05 for “Ignore” decisions, in the case of centralized authority. These probabilities range between-10% and +20% depending on UAI and PDI scores’ relative differences to American scores.

(3) Tolerance of waiting time for decisions: This matrix defines how long an

actor waits for a decision before it assumes delegation by default and makes its own

decision. Actors in different managerial roles have different time-out durations. The

power distance index (PDI) and individualism index (IND) are linked to wait times,

because of the high tolerance for inequality (Hofstede, 1991, p.37, table2.3) and high

value placed on group harmony (Hofstede, 1991, p73).

Preposition 5: High PDI countries show longer tolerance in waiting for a

decision.

Preposition 6: High IND countries show less tolerance in waiting for a decision.

This assumes that both PDI and IND equally affect magnitude of tolerance in

waiting for a decision.

74

Distance X = ((PDI – American PDI) – (IND – American IND))/2

Table 5.5 shows the adjustment factors of wait duration for decision making. The

large X value (High PDI and Low IND) represents high tolerance to wait for decision

making. The deviated range is set between -20% and + 40% from the defaulted duration.

Figure 5.4 indicates the adjusted ranges for Time-To-Wait-For-Decision-Making based

on the adjustment factors.

Table 5.5: Adjustment Factor for Time-to-Wait-for-Decision-Making

Adjustment factor X > 40 40 > X > 20

(Pattern J)

20 > X> -20

(Pattern A)

-20 > X

Duration of time to wait for

decision making (PM, SL, ST)

+40% +20% 0% -20%

0

200

400

600

800

1000

1200

1400

1600

X > 40

40 > X > 20

20 > X> -20

-20 > X

Defaulted wait duration

Adjusted ranges of wait duration by PDI and IND scores

PM SL, ST

Figure 5.4: Adjusted Duration for Time-to-Wait-for-Decision-Making

Note: Table 5.5 illustrates the adjustment factors of wait duration for decision making. The large X value (High PDI and Low IND) represents high tolerance to wait for decision making. The deviated range is set between -20% and + 40% from the defaulted duration. Figure 5.4 shows the adjusted wait duration for decision making.

(4) Attendance to communication: In the VDT model, when a participant

picks up a communication item, it has to decide whether to attend to the communication.

This behavior defines the probability that an actor attends to a given type of

75

communication, based on the matrix level of the organization, the uncertainty avoidance

index (UAI), and the individualism index (IND). UAI is positively correlated to the

probability of attendance to certain types of communication, because high UAI countries

tend to acquire current and precise information by attending to communications. IND is

negatively related to the probability of attendance to communications, because of

different attitudes toward group communication.

Preposition 7: High UAI countries show higher attendance ratios to

communications.

Preposition 8: High IND countries show lower attendance ratios to

communications.

UAI is positively correlated to the probability of attendance to certain types of

communication, because high UAI countries tend to acquire current and precise

information by attending to communications. IND is negatively correlated to the

probability of attendance to communications, because of different attitudes toward group

communication.

AdjustedAttendanceProbability = OriginalProbability * Adjustment Factor

The Adjustment Factor is calculated by Table 5.6

Distance Y = ((UAI – American UAI) – (IND – American IND))/2

Table 5.6 illustrates the adjustment factors of attendance probability to

communication. A large Y value (High UAI and Low IND) represents high probability

to attend both formal and informal communications. The range of deviation is set

between -10% and + 20% from the defaulted frequency. This assumes no effect on

“noise” frequency, since “noise” is an environmentally contingent, rather than a cultural,

attribute of a project. Figure 5.5 illustrates the adjusted attendance probabilities to

communication.

76

Table 5.6: Adjustment Factor for Attendance Probability to Communication

Adjustment factors Y > 40

(Pattern J)

40 > Y > 20 20 > Y> -20

(Pattern A)

-20 > Y

Attendance of Communication

(Meeting Communication)

+20% +10% 0% -10%

Attendance of Communication

0

0.2

0.4

0.6

0.8

1

H M L

Matrix Level

Comm

Meet

Adjusted ranges of attendance probabilities to communication

Defaulted attendance probabilities to communication

Figure 5.5: Adjusted Probabilities for Attendance Probability to Communication

Note: Table 5.6 shows the adjustment factors of attendance probability to communication. A large Y value (High UAI and Low IND) represents high probability to attend both formal and informal communications. The range of deviation is set between -10% and + 20% from the defaulted frequency. This assumno effect on “noise” frequency, since “noise” is an environmentally contingent, rthan a cultural, attribute of a project. Figure 5.5 indicates the adjusted attenprobabilities to communication. The probabilities depend on the level of matrix strength.

es ather

dance

(5) Response of communication probability: In the VDT model, the users set

the information exchange frequency and initial noise frequencies, which are then adjusted

depending on the level of formalization and individualism index (IND). We start with the

assumption that a high level of formalization in an organization enables it to reduce the

frequency of communication, because of standardized process and rules. Hofstede

argues that the degree of formalization can be predicted based on the PDI – UAI

dimensions (Hofstede, 1991, p.152). High PDI and high UAI countries, with full

bureaucracies, tend to standardize the work process rather than using mutual adjustment.

77

This means that high PDI and high UAI countries are most likely to have a highly

formalized coordination system, while low PDI and low UAI countries prefer to

standardize outputs. Thus, each nation has a preference for the formalization level based

on their PDI and UAI dimensions.

Preposition 9: High PDI and high UAI countries show a higher formalization

level.

Table 5.7: Adjustment Factor for Response Probability

Adjusted value Z > 40

(Pattern J)

40 > Z > 20 20 > Z> -20

(Pattern A)

-20 > Z

Response Volume

(InfoExchange, Meeting)

+20% +10% 0% -10%

0

0.5

1

1.5

2

2.5

H M L

Comm

Meet

Defaulted response probabilities

Adjusted response probabilities

Formalization Level

Figure 5.6: Adjusted Probabilities for Response Probability

Note: Table 5.7 shows the adjustment factors for probability of attending to communication. Figure 5.6 illustrates the adjusted response probabilities. The larger Z value (High UAI and High PDI) represents high response volume for both formal and informal communications. The range of deviation is set between -10% and + 20%. This assumes no effect on “noise” volume, since “noise” is not a cultural attribute. The response volume depends on the level of formalization.

High PDI and high UAI countries show a higher formalization level, increasing

the frequency of formal meeting and informal information exchange.

78

AdjustedResponseProbability = OriginalProbability * Adjustment Factor

Adjustment Factor is calculated by Table 5.7

Distance Z = ((PDI – American PDI) – (UAI – American UAI))/2

The larger Z value (High UAI and High PDI) represents high response probability

for both formal and informal communications. The deviation is set between -10% and +

20%. This research assumes that cultural factors have no effect on “noise” volume, since

“noise” is not a cultural attribute.

(6) Information demand for communication: In the VDT model, the

communication volume, or information demand, is pre-set by the user as a global

information exchange frequency. The uncertainty avoidance index (UAI) affects this

information demand for communication. As discussed in chapters 2, high UAI countries

prefer “precise” and “quality” jobs. In order to achieve a precise” and high quality job,

team members exchange information more frequently in order to minimize

miscommunication or misunderstanding. In other words, high UAI countries show high

information demand for communication.

Preposition 10: High UAI countries show greater information demand for

communications among team members.

Table 5.8: Adjustment Factor for Probability-of-Attending-to-Communication

Adjustment factor UAI > 40

(Pattern J)

40 > UAI > 20 20 > UAI> -20

(Pattern A)

-20 > UAI

Message Demand for

Communication

(PM, SL, ST)

+20% +10% 0% -10%

Note: Table 5.8 shows the adjustment factors for information demand for communication. A larger UAI value represents greater demand for information exchange. The deviation is set between -10% and + 20%. This assumes no difference in roles of actors (PM, SL, ST)

79

Table 5.8 shows the adjustment factors for information demand for

communication. A larger UAI value represents greater demand for information exchange.

The deviation is set between -10% and + 20%. This assumes no difference in roles of

actors (PM, SL, or ST)

Table 5.9 summarizes two sets of micro-level behaviors determined by the above

prepositions. The six behavior parameters comprise the individual behavior patterns,

Japanese and American. The American behavior pattern has average in every parameter,

since the default parameters are assumed to represent the American behavior pattern. On

the other hand, the Japanese behavior pattern is adjusted based on cultural value gaps.

Table 5.9: Sets of Micro-Level Behavior

Decision Policy Pattern J Pattern A

Decision Maker Policy Centralized Average

Wait Time for Decision Making Long Average

Decision Type Less rework Average

Communication Policy Pattern J Pattern A

Attendance to communication Higher attendance Average

Information demand for communication Increased Average

Response of communication volume Increased Average

Note: Table 5.9 illustrates sets of micro-level behavior patterns. A micro-level behavior pattern is the defaulted version of micro-level behavior in VDT since VDT has been calibrated based on American firms and people. On the other hand, the J micro-level behavior pattern shows the shift relative to the A behavior pattern.

5.2.3 Task Complexity

(1) Task dependencies: To emphasize the impact of culture on team

performance, contingent on task complexity, task dependencies are varied in this research.

Task dependencies can be classified into four arrangements (Thompson, 1967; Bells and

Kozlowski, 2002): a pooled workflow, a sequential workflow, a reciprocal workflow, and

an intensive workflow. These four workflows represent basic kinds of work processes in

real projects.

80

A pooled workflow is a structure in which tasks and activities are performed

separately by all team members and then combined into a finished product. This is

similar to the concept of “fit” as one of the basic types of dependencies (Malone et al,

1999; Zlotkin, 1995). In a case of the construction industry, to construct several

residential buildings in a district can be considered as an example. The residential

buildings do not need any coordination activities (except for a fence work between land

boundaries) among them, implying a pooled type workflow. Generally speaking, pooled

workflow implies low task complexity since information exchanges and rework

relationships among tasks are low level in comparison with other types of workflow.

Task1

Task2

Task3

Task4

Figure 5.7: A Pooled Workflow Structure

A sequential structure is a workflow in which tasks and activities flow

sequentially from one to the next. This structure represents the design-bid-build project

in this research. The design-bid-build procedure is still predominant in the construction

industry. Traditionally, field construction is not begun until the architect-engineer has

completed and finalized the design (Clough et al, 2000). The sequential procedure

generally imposes low to medium complexity, since information exchanges and rework

relationships among tasks are limited to hand-offs between consecutive tasks.

Task4Task3Task2Task1

Figure 5.8: A Sequential Workflow Structure

A reciprocal structure is the minimum unit of interdependent workflow that lies

between sequential and intensive workflows. This structure represents Design&Build

projects in this research. At the Design&Build projects, the construction phase starts

81

before the design phase completes, requiring frequent reciprocal information exchanges

among tasks such as architect design, structural design, Mechanical/Electrical/Plumbing

(MEP) design, and construction. The reciprocal procedure usually imposes medium to

high complexity, because of high information exchanges and rework relationships among

tasks.

Task 1

Task 2

Task3

Task4

Figure 5.9: A Reciprocal Workflow Structure

An intensive structure is the most interdependent workflow. It represents the fast-

track project. The fast-track procedure refers to the overlapping of project design and

construction. As the design of progressive phases of the work is finalized, these designs

are put under contract, a process commonly referred to as “phased construction.”

Construction of the early phases of the project is begun while later stages are still on the

drawing boards. This procedure of overlapping the design and construction phases can

appreciably reduce the total time required to achieve project completion (Clough et al,

2000). The requirements of many current projects are increasing task dependencies on

each other. For instance, NASA, the Jet Propulsion Laboratory (JPL) Advanced Project

Development Team, use the Integrated Concurrent Engineering (ICE) approach to

conduct highly overlapped task flows and highly interdependent tasks by multiple

collocated multidisciplinary teams (Chachere et al, 2004).

Task Task

Task Task

Figure 5.10: An Intensive Workflow Structure

82

Table 5.10 shows a summary of workflows. The level of complexity is shifted

from low to high, corresponding to the levels of task interdependencies.

Table 5.10: Summary of Workflows

Task Workflow

Interdependencies Pooled type Sequential type Reciprocal type Intensive type

Complexity Low High

*Adapted from Thompson (1967), and Bell and Kozlowski (2002), “A typology of virtual

teams”

(2) Project intensity: In the VDT model, there are three settings that

determine project intensity: information exchange ratio, project error probability, and

functional error probability. Project intensity obviously depends on the project type. In

this research, project intensity is determined in conjunction with four prototypical project

workflows: pooled, sequential, reciprocal, and intensive.

Table 5.11: Setting of Project Intensity

Pooled

Workflow

Sequential

Workflow

Reciprocal

Workflow

Intensive

Workflow

Information Exchange Ratio (IE) Low

(0.4)

Med

(0.47)

Med

(0.53)

High

(0.6)

Project Error Probability (P) Low

(0.08)

Med

(0.09)

Med

(0.11)

High

(0.12)

Functional Error Probability (F) Med

(0.1)

Med

(0.1)

Med

(0.1)

Med

(0.1)

Information exchange probability measures the amount of communication in

the project between positions that are responsible for tasks linked by communications

links. The information exchange probability is set for the project as a whole (SimVision

83

4.0.0 Help Files). In the case of the pooled workflow, since actors conduct tasks and

activities separately, the information exchange ratio can be set to a low value. On the

other hand, in the case of the intensive workflow, since task dependencies are

complicated, actors need to acquire relevant task information as much as they can. Thus,

the information exchange ratio tends to be high (Table 5.11).

Project-error probability is typically set in the range between 0.05 (low) and

0.15 (high). Project error probability can be set to a low value if the project involves

relatively standard tasks and routine work processes. It should be set to a higher value

for nonstandard tasks and innovative work processes (SimVision 4.0.0 Help Files). Since

construction tasks are relatively non-routine work processes, it is fair to set a common

value (0.10) as the average of the four workflows. The intensive project type generally is

of short duration with high work volume, and thus its project error probability is

potentially high, 0.12 (Table 5.11).

Functional-error probability is typically set in the range from 0.05 to 0.15. The

functional-error probability can be set with a low value if the project involves relatively

well-understood technology and standard work processes. This probability can be set to a

high value if the project involves unproven technology or innovative work processes

(SimVision 4.0.0 Help Files). Since technological level is out of the scope of this

research, all three prototypes are assumed to have a medium error probability, 0.10

(Table 5.11).

5.2.4 Team Experience

Team experience refers to previous experience working together in similar type of

projects. Based on our observations, in three out of four projects, team members have no

previous experience working together. Generally speaking, in the construction industry,

team members tend to be assembled on a project basis, and dissolved when the job is

done. In the case of international projects, this discrete tendency is enhanced because of

different location project by project, unstable international construction market, and

numerous possible combinations of main players. This inconsistency implies that team

experience is most likely to be low. However, there are a few cases where team cohesion

was achieved by overcoming economic instability. In this case, cultural conflicts and

84

misunderstandings were mitigated and vanished. Thus, in this research, team experience

is varied from low to high as an independent variable. The literature on cultural diversity

suggests that managing cultural diversity has benefits, such as cost effectiveness,

creativity, and problem-solving quality (Taylor & Stacy, 1991). This implies that a

mixed-cultural team with “high” team experience can potentially have better team

performance. However, in this research, there is not enough evidence to indicate positive

effects correlated to team experience.

Table 5.12: Team Experience

Team Experience Low Med High

5.3 Experimentation Figure 5.11 shows the framework of an intellective experiment. Independent variables

for the intellective experiment are organization styles, micro-level behavior patterns,

project complexity and team situation. The intellective experiment simulated a total of

48 scenarios (= 2 organization styles x 2 micro-behavior patterns x 4 task complexity

levels x 3 team situation levels). Dependent variables of a simulation are project

duration, cost and quality risks (Figure 5.11).

For experimental purposes, the actor and task configurations are identical13. The

VDT model is designed to predict duration, cost, quality risks and project risks as

measures of team performance. The VDT model displayed the simulated and the critical

path method (CPM) duration. The simulated duration is calculated by considering the

simulated work volume and workflow. The CPM duration is calculated by considering

the designed work volume and workflow. The gap between the simulated and the

designed work volume is called “hidden work” (Levitt & Kunz, 2002), and is caused by

rework, coordination efforts, and wait time for decisions. Thus, this hidden work

represents the efficiency of team performance. The cultural practice and values

differences in the Japanese vs. American structures and micro-behaviors of the actors 13 Actor and task configurations include actors’ skills, skills required by tasks, duration of tasks, hourly salary of actors, total number of team participants (all team is composed of 7 members including 1 project manager, 2 sub team leaders, and 4 sub team members), and tasks’ responsibility position.

85

cause differences in “hidden work volume (Levitt & Kunz, 2002)”, and hence in schedule,

cost and quality outcomes. Thus, we analyze three dependent variables, - 1) hidden work

volume, 2) product quality risks14, and 3) project quality risks15 -, to analyze the impacts

of changes in organization styles and micro-level behaviors on team performance.

Figure 5.11: Framework of Intellective Experiments

Note: This figure illustrates the framework of an intellective experiment. There are four independent variables: (1) organization style, (2) micro-level behavior, (3) task complexity, and (4) team experience.

Based on observed differences in cultural practices, there are two sets of organization styles: American and Japanese organization styles (1). Similarly, with observed differences in cultural values, there are two sets of micro-level behavior patterns: American and Japanese individual behavior patterns (2). This intellective experiment tests four different levels of task complexities such as pooled, sequential, reciprocal and intense workflows (3). Additionally, I examine three different levels of team experience: low, medium and high team experience (4).

The intellective experiment simulates a total of 48 scenarios (= 2 organization styles x 2 micro-behavior patterns x 4 task complexity levels x 3 team situation levels). Dependent variables of this simulation are project duration, cost, and quality risks.

14 Product quality risk index represents the likelihood that specialized components produced by this project will have defects based on rework and exception handling (Help function in the SimVision ®). 15 Project quality risk index represents the likelihood that the overall product produced by this project will not be integrated at the end of the project, or that the integration will have defects based on rework and exception handling (Help function in the SimVision ®).

86

5.3.1 Settings for Intellective Experiment

For the simulation experiments, the detailed settings of the VDT model are as follows.

Note: These figures illustrate examples of As shown in Figures 5.12 and 5.13, both orand required work volume as the intense coseven members, including one project manteam members. We change either structurepossess. shows precedence links among shows rework and communicat shows work assignment betwee

(1) Simulation Engine: SimVision

(2) Actors: There are seven actors

has a role such as one project manager (PM)

team members (ST). Each actor has a profil

profession, and work experience. In this inte

are identical: generic profession, medium sk

Additionally, all team members have either J

With the limitation of the current VDT mode

mono-cultural cases such as comparing Japa

cultural teams.

16 SimVision-R was developed by Vité CorpoAustin Texas.

Figure 5.13: Example of Japanese Organization Structure Type with IntensiveComplexity

Figure 5.12: Example of American Organization Structure Type with IntensiveComplexity

the intense coordination complexity cases. ganizations have the exactly same workflow mplexity cases. All teams are composed of

ager, two sub-team leaders, and four sub- types or micro-level behavior patters actors

tasks ion links among tasks n team members and tasks

-R16 – Educational Version 3.11.1

in the idealized organization. Each actor

, two sub team leaders (SL), and four sub

e such as professions, skill level of each

llective experiment, all of actors’ profiles

ill level, and medium work experience.

apanese or American behavior patterns.

l, this intellective experiment examines

nese teams vs. American teams, not mixed

87

ration, and is licensed from ePM, LLC,

(3) Tasks: There also are seven tasks. Each task has a responsible actor. Each

task is identical and has 120 person days work volume. The total work volume is 840

person days, which represents a medium sized construction project.

(4) Rework and Communication links: Rework and communication links are

ascribed according to the task intensity. Experiments to find an organizational equivalent

of the Reynolds Number in Fluid Mechanics (Fyall, 2001) refers to a “turbulent point” at

which hidden work volume increases dramatically, as a function of rework links. Thus,

the amount of rework and communication versus the total number of tasks were

examined as part of task complexity.

Table 5.13: Rework and Communication Ratio*

Pooled Sequential Reciprocal Intensive

Rework ratio 0 0.4 1 1.3

Communication ratio 0.9 0.9 0.9 1.7

*Ratio = number of rework or communication items / total number of tasks

(5) Parameters (Appendix A and Appendix B): Each nation has its own set of

micro-level behaviors. Behavior parameters are set based on observations and existing

theories. Each set of behavior patterns is shown in Appendix A and Appendix B.

5.3.2 Simulated Results

Table 5.14 shows the summary of the simulated results, including: duration, hidden work

volume, cost, functional quality risks, and project quality risks.

Table 5.14: Summary of Simulated Results (person-months)

88

Complexity Low High

Workflow Pooled Sequential Reciprocal Intensive

Structure Type J Type A Type J Type A Type J Type A Type J TypeA

Duration (Critical Path Method)

Duration (M) 8.00 8.10 29.6 28.8 30.7 29.6 13.5 13.1

Standard dev (0.18) (0.15) (1.30) (0.80) (1.60) (1.00) (1.10) (0.60)

Two sample

tests (n=100) Type J = Type A Type J = Type A Type J = Type A Type J = Type A

Hidden Work Volume

Hidden Work

Volume (M) 3.535 4.455 14.595 12.570 26.020 21.125 29.185 38.150

Two sample

tests (n=100) Type J < Type A Type J > Type A Type J > Type A Type J < Type A

Cost

Cost ($1,000) 281 288 355 343 431 401 446 497

Standard dev (2.65) (2.78) (27.49) (17.16) (47.60) (33.89) (33.41) (56.28)

Two sample

tests (n=100) Type J < Type A Type J > Type A Type J > Type A Type J < Type A

Functional (Product) Quality Risks

Quality Index

(FRI) 0.469 0.468 0.466 0.464 0.467 0.461 0.478 0.48

Standard dev (0.044) (0.037) (0.037) (0.041) (0.035) (0.034) (0.033) (0.022)

Two sample

tests (n=100) Type J = Type A Type J = Type A Type J = Type A Type J = Type A

Project Quality Risk

Project Risk

Index (PRI) - - 0.267 0.437 0.284 0.467 0.279 0.472

Standard dev. - - (0.044) (0.067) (0.037) (0.046) (0.031) (0.033)

Two sample

tests (n=100) - Type J < Type A Type J < Type A Type J < Type A

Note: (1) Total simulated work volume is the sum of production work volume and coordination work volume (Jin and Levitt, 1996, pp175)

Hidden Work Volume = Total Simulated Work Volume – Designed Work Volume (2) For each scenario, we run 100 trials and calculate means and standard deviations. (3) Product quality risk represents the likelihood that components produced by the project have defects based on rework and exception handling (Jin and Levitt 1996, pp179) (4) Project quality represents the likelihood that the components produced by the project will not be integrated at the end of the project, or that the integration will have defects based on rework and exception handling (Jin and Levitt, 1996, pp179).

89

(1) Project Duration: The simulated project duration is the total duration of

activities that are on the critical path to finish a project. As can be seen in Table 5.14,

there is no significant difference between Japanese and American structural styles.

Similar range of standard deviation suggests that the duration based on simulated CPM

duration is not significantly affected by changes in structural styles.

(2) Hidden work volume: Hidden work volume represents the efficiency of

team performance. The larger the hidden work volume becomes, the worse the efficiency

of team performance in cost, schedule, and quality risks will be. In a VDT sense, hidden

work volume is the barometer of potential risks of a project.

- Effects of task complexity: As seen in Table 5.14, the hidden work volume

is linearly, rather exponentially, increased as task complexity is shifted from low to high.

This simulated result shows a consistent tendency with Galbraith (Galbraith, 1974), who

argued that “the greater the uncertainty of the task, the greater the amount of information

that has to be processed between decision makers during task execution in order to

achieve a give level of performance”. This implies that the experiment settings, and the

idealized task complexity, are appropriate and robust.

Effects of Changes in Org. Styles

0.0

15.0

30.0

45.0

Pooled Sequencial Reciprocal Intensive

Project Complexity

Hid

den W

ork

Volu

me

J Org. style

A Org. style

American Style Decentralized authority Medium formalization

Flat hierarchy

23%

The effects of Changes Japanese Style

Centralized authority High formalization

Multi-level hierarchy

Figure 5.14: Effects of Changes in Organizational Structure Types

Note: This figure compares the performance of Japanese vs. American organization structure types. The X axis shows the level of project complexity. The Y axis shows total hidden work volume in person-months. Task interdependencies such as pooled, sequential, reciprocal, and intensive workflow represent a range from low to high task complexity respectively. The defaulted version of micro-level behaviors is used for the two cases. Black lines represent standard deviations (Table 5.14).

90

- Effects of changes in the structural styles: When comparing hidden work

volume between Japanese (J) and American (A) structural styles, the Japanese structure

style shows better performance when the task complexity is pooled (low) and intense

(high) cases. On the other hand, the American structural style shows better performance

in the cases of medium task complexity (sequential and reciprocal cases). For cost

outcomes, this tendency is exactly the same as for hidden work volume. Figure 5.14

shows the more detailed comparison data by considering the level of project complexity.

A key result is that the American structure style has better performance in the cases of

medium task complexity than that of the Japanese structure style. The Japanese structure

style has better performance in the case of high task complexity than that of the American

structure style. The largest magnitude of the impacts of changes in organization styles is

23% in the case of intensive project workflow.

- Effects of changes in team experience: Table 5.15 shows the effects of

changes in team experience from low to high. The basic tendency is that the better the

team experience, the better the team performance will be (=lower hidden work volume).

The other finding is that the American structural style is more sensitive to changes of

team experience, in particular in the case of intensive workflow. When considering the

decentralized strategy, every team member needs to work effectively independently. If

one unskilled employee joins a team, it may devastate the whole group’s performance.

Thus, American teams generally may tend to adopt a strategy like “hire the right

<person> in the right position (Ouchi, 1981: Nakane, 1970)”.

Table 5.15: Effects of Team Experience

Team Experience J organization style

High Medium Low

Differences

Low-High

Pooled 3.30 3.54 4.01 0.71

Sequential 14.21 14.60 14.82 0.61

Reciprocal 24.79 26.02 25.69 0.90

Intensive 25.95 29.19 34.70 8.75

91

Team Experience A organization style

High Medium Low

Differences

Low-High

Pooled 4.03 4.46 5.31 1.28

Sequential 11.85 12.57 14.18 2.33

Reciprocal 20.83 21.13 22.41 1.58

Intensive 26.40 38.15 73.91 47.51

- Effects of changes in behavior patterns: Figures 5.15 and 5.16 show the

effects of changes in micro-level behavior patterns on hidden work volume. American

behavior cases with the American organization style show better performance than

Japanese behavior cases with the American organization style (Figure 5.15). In particular,

the intensive workflow has 29% differences in performance between Japanese and

American behavior cases. On the other hand, Japanese behavior cases with the Japanese

organization style show better performance than American behavior cases with the

Japanese organization style (Figure 5.16). These simulation results confirm Hofstede’s

proposition that “each culture has its preferred coordination mechanism” (Hofstede,

1991). Specifically, organizational performance of workers who have the culture’s

preferred micro-level behavior is positively correlated to the use of each culture’s typical

organization style, in cases of medium to high task complexity. In the case of pooled and

sequential workflow, the differences between Japanese and American behavior patterns

are relatively small. This implies that increasing task complexity amplifies the impact of

cultural practice vs. behavior mismatches, as we would expect, since it increases the

frequency of exceptions that will arise in executing direct tasks (Galbraith 1973). The

effect of changes in micro-level behavior patterns with the Japanese organization style

(5%) is smaller than with American organization style (29%). This implies that the

centralized organization style such as the Japanese organization style has less impact of

changes in micro-level behavior patterns.

92

Effects of Changes in Behavior Pattern with AM Org. Style

0.0

15.0

30.0

45.0

60.0

Pooled Sequencial Reciprocal Intensive

Project Complexity

Hid

den

Work

Volu

me

J behavior pattern

A behavior pattern

American Style

Org

. Sty

le

Decentralized authority Medium formalization

Flat hierarchy

29%

American pattern

Beh

avio

r pat

tern

(a) Individual decision making

(b) Individually-based communication

Japanese pattern (a) Consensual decision making (b) Group-based communication

Figure 5.15: Effects of American vs. Japanese Micro-Level Behavior Patterns with American Organizational Structure Type

Note: This compares the performance of Japanese vs. American micro-level behavior patterns for the American organization style. The X axis shows the level of task workflow such as pooled, sequential, reciprocal, and intensive interdependencies. Each workflow represents from low to high task complexity respectively. The Y axis represents total hidden work volume in person-months.

Effects of Changes in Behavior Patterns With JP Org. Style

0.0

15.0

30.0

45.0

60.0

Pooled Sequencial Reciprocal Intensive

Project Complexity

Hid

den W

ork

Volu

me

J behavior pattern

A behavior pattern

Japanese Style

Org

. Sty

le

Centralized authority High formalization

Multi-level hierarchy

American pattern

Beh

avio

r pat

tern

5% (a) Individual decision making

(b) Individually-based communication

Japanese pattern (a) Consensual decision making (b) Group-based communication

Figure 5.16: Effects of American vs. Japanese Micro-Level Behavior Patterns with Japanese Organizational Structure Type

Note: This compares the performance of Japanese vs. American micro-level behavior patterns for the Japanese organization style. The X axis shows the level of task workflow such as pooled, sequential, reciprocal, and intensive interdependencies. Each workflow represents from low to high task complexity respectively. The Y axis represents total hidden work volume in person-months.

93

(3) Cost: As seen in Table 5.14, the cost results indicates that Japanese teams

are more efficient in the case of high task complexity (intense case), while American

teams are more efficient in the cases of medium task complexity (sequential and

reciprocal). In the case of low task complexity, the Japanese team is better than the

American. However, differences between the two are subtle, and in the range of one

standard deviation.

(4) Quality Risks: Table 5.14 shows the differences in the function-quality-

risk and project-quality-risk indexes. In the VDT model, the functional (product)-

quality-risk index represents the likelihood that components produced by this project

have defects based on rework and exception handling (Help function in the SimVision ®).

Project quality risk index represents the likelihood that the components produced by this

project will not be integrated at the end of the project, or that the integration will have

defects based on rework and exception handling (Help function in the SimVision ®).

Any quality risk below 0.5 is acceptable. If the risk is greater than 0.5, it indicates high

product quality risk (SimVision 4.0.0 Help Files).

- Product quality risks: Table 5.14 and Figure 5.17 show product quality risks

of both Japanese and American cases. There is no significant difference between

Japanese and American cases, ranging between 0.45 and 0.5. All the cases are below 0.5,

meaning acceptable range in defect risks.

Effects on Quality Risk Index

0.00

0.10

0.20

0.30

0.40

0.50

Pooled Sequencial Reciprocal Intensive

Project Complexity

Prod

uct Q

uality

Inde

x

J Org. style

A Org. Style

Figure 5.17: Effects on Product Quality Risk

94

- Project quality risks: Table 5.14 and Figure 5.18 shows effects on project

quality risks — the risks of failed integration among components of a product. Japanese

cases obviously show significantly lower risk, ranging between 0.25 and 0.3, than that of

American cases, ranging between 0.45 and 0.50. This implies that the Japanese cases

embody a relatively conservative strategy, because they fall far below the threshold value

set from experience at 0.50. On the other hand, American cases can be seen as a

relatively aggressive strategy, because they are close to the 0.50 risk threshold. Which

strategy is better depends on people’s preference. Based on Hofstede’s dimension, UAI

index can be linked to this. High UAI countries may prefer to take Japanese type strategy,

because they are most likely to show high anxiety level about uncertainty. On the other

hand, in the weak uncertainty avoidance countries, people may prefer to take American

type strategy, because of low anxiety level of uncertainty.

Effects on Project Risk Index

0.00

0.10

0.20

0.30

0.40

0.50

Pooled Sequencial Reciprocal Intensive

Project Complexity

Proj

ect R

isk

Inde

x

J Org. styleA Org. Style

StDev: 0.04

StDev: 0.03

Figure 5.18: Effects on Project Quality Risks

5.3.3 Components of Hidden Work

Hidden work volume is composed of three elements: rework, coordination efforts, and

wait time for decisions. Qualitative tendencies in each case, for each component of

hidden work, can be captured using the Japanese/American (J/A) effective index.

95

J/A effective index

Hidden work volume of a J behavior case

Hidden work volume of an A behavior case =

(1) Rework is defined as the work that has to be redone on a task due to

exceptions that occur in another task linked to it by a rework link. Rework occurs in the

dependent task, or the task at the arrow end of the rework link (SimVision 4.0.0 Help

Files). As Table 5.10 shows, the index is below 1.0 in almost all cases. This means that

Japanese actors do less rework than American actors in all cases of task complexity and

organization structure. Thus, the Japanese micro-level behavior pattern consistently

shows less rework than the American micro-level behavior pattern.

This tendency is explained by the VDT mechanisms and the encoded micro-level of

behaviors. In VDT, there is a Verification Failure Probability (VFP) that refers to the

probability of failure. For instance, a high VFP means a high probability of failure, while

a low VFP means a low probability of failure. The program VFP is initially set when a

modeler set the functional and project error probabilities. In the intellective experiment,

functional and project error probabilities are set based on project complexity (Chapter

5.2.3; Table 5.11), implying that these initial probabilities do not make any differences in

performance between Japanese and American behavior cases. During a simulation, the

initial VFP is adjusted based on two factors: skill mismatches17 and decision types.

Since the required skill of tasks and actors’ skill level are identical, skill

mismatches are neglected.

Decision type means the type of decisions (rework, correct, or ignore) that are

made about an exception in a task. The initial VFP is adjusted depends on the response

that an exception is ignored, corrected, or reworked. The VFP increases if an exception

is ignored and decreases if it is reworked. The encoded micro-level behavior parameters

— decision making policy and decision type — affect the probability that an exception is

either ignored or reworked. In Decision Making Policy, for instance, high PDI score in

17 Skill mismatches mean that the position responsible for a task does not have the skill required by the task, doesn't have it at the right level, or has a lower application experience setting than required by the task. Skill mismatches have the most damaging effect on a VFP.

96

Japan increases the probabilities that a project manager make a decision for an exception.

A behavior of PM is to make more “correct” and “rework” decisions, but less “ignore”

decisions. Additionally, high UAI score in Japan increases the probability of “rework”

decisions and decreases the probability of “ignore” decisions. Therefore, the Japanese

micro-level behavior pattern that increases “correct” and “rework” decisions and

decreases “ignore” decisions on an exception, substantially decreasing the initial VFP.

Decreased VFP gives a reason of why the Japanese micro-level behavior pattern shows

consistently less rework than the American micro-level behavior pattern.

Table 5.16: Comparison of Reworking Volume

Team Experience J / A effective index

(J behavior vs. A behavior cases) Low Medium High

Pooled 0.99 1.00 1.00

Sequential 0.94 0.94 0.95

Reciprocal 0.94 0.93 0.95 J structure style

Intensive 0.94 0.93 0.95

Pooled 0.99 1.02 0.99

Sequential 0.98 0.97 0.96

Reciprocal 0.98 0.96 0.97 A structure style

Intensive 1.00 0.97 0.94

i

Note: This table compares the rework volume of Japanese vs. American micro-level behavior patterns for each structure type. I use the J/A effective index as a measurement. The J/A effective index is calculated as follows:

J/A effective index

Hidden work volume of a J behavior case

Hidden work volume of an A behavior case =

The greater hidden work volume becomes, the less efficient a project team will be. If J/A effective index < 1.0: Performance of a J behavior case is better than that of an American behavior case. If J/A effective index =1.0: Performance of a Japanese behavior case is equal to that of an American behavior case. If J/A effectiveindex < 1.0: Performance of a American behavior case is better than that of a Japanese behav

or case.

97

Organization styles slightly affect rework volume. High level of centralization as

a component of the Japanese organization style increases the probabilities that the highest

position make a decision on an exception. Decentralized policy (low level of

centralization) as the American organization style affect oppositely. However,

organizational differences, high centralization vs. low centralization, are not clearly

determined in Table 5.16.

(2) Coordination models the information flow among positions. Coordination

volume is measured as the sum of two different types of communication: one-to-one

information exchange between positions or persons, and group meetings (SimVision

4.0.0 Help Files). The simulated results for coordination efforts show that all cases are

above 1.0. This means that American people coordinate efficiently in all cases. Thus,

the American micro-level behavior pattern consistently shows efficient coordination in

comparison with the Japanese micro-level behavior pattern.

This tendency is explained by the encoded micro-level behavior parameter,

attendance probability to communications. High UAI and Low IND in the Japanese

behavior pattern represent high probability to attend both formal and informal

communications, increasing the probabilities of information exchange between positions

or persons, and group meetings. Therefore, Japanese people need to spend more time for

coordination than American people do.

98

Table 5.17: Comparison in Coordination

Team Experience J / A effective index

(J behavior vs. A behavior cases) Low Medium High

Pooled 1.13 1.15 1.14

Sequential 1.08 1.10 1.10

Reciprocal 1.08 1.10 1.11 J structure style

Intensive 1.11 1.12 1.13

Pooled 1.15 1.16 1.14

Sequential 1.11 1.13 1.12

Reciprocal 1.12 1.14 1.14 A structure style

Intensive 1.21 1.17 1.14

Note: This table compares the coordination efforts of Japanese vs. American micro-level behavior patterns for each structure type. I use the J/A effective index as a measurement. The J/A effective index is calculated as follows:

J/A effective index

Hidden work volume of a J behavior case

Hidden work volume of an A behavior case =

The greater hidden work volume becomes, the less efficient a project team willbe.If J/A effective index < 1.0: Performance of a Japanese behavior case is better thanthat of an American behavior case. If J/A effective index =1.0: Performance of a Japanese behavior case is equal to that of an American behavior case. If J/A effective index < 1.0: Performance of a American behavior case is better than that of a Japanesebehavior case.

(3) Time to wait for decision making arises when positions report exceptions

to supervisors and wait for their supervisors to make decisions about how to deal with the

exceptions. The components include the time a position waits for a response from the

supervisor about how to handle an exception, plus any time the position waits for

exception resolution before making the decision (SimVision 4.0.0 Help Files). The index

for wait time is above 1.0 in all cases. This means that Japanese actors are most likely to

wait longer for decision making than American actors. Thus, the American micro-level

behavior pattern consistently shows efficient decision making in comparison with the

Japanese micro-level behavior pattern.

99

Table 5.18: Comparison of Time-to-Wait-for-Decision-Making

Team Experience J / A effective index

(J behavior vs. A behavior cases) Low Medium High

Pooled 1.01 1.09 1.02

Sequential 1.07 1.07 1.10

Reciprocal 1.06 1.05 1.09 J structure style

Intensive 1.06 1.04 1.09

Pooled 1.02 1.12 1.05

Sequential 1.01 1.04 1.03

Reciprocal 1.05 1.03 1.04 A structure style

Intensive 1.06 1.02 1.02

Note: This table compares the wait-time-for-decisions of Japanese vs. American micro-level behavior patterns for each structure type. I use the J/A effective indemeasurement. The J/A effective index is calculated as follows:

x as a

J/A effective index

Hidden work volume of a J behavior case

Hidden work volume of an A behavior case =

The greater hidden work volume becomes, the less efficient a project team will be. If J/A effective index < 1.0: Performance of a Japanese behavior case is better than that of an American behavior case. If J/A effective index =1.0: Performance Japanese behavior case is equal to that of an American behavior case. If J/A effective index < 1.0: Performance of a American behavior case is better than that of a Japanese behavior case.

of a

This tendency is explained by the micro-level behavior parameter, time-to-wait-

for-decision-making. High PDI and Low IND represent high tolerance to wait for

decision making. Therefore, Japanese people are most likely to wait longer time for their

super-ordinate’ decisions.

In centralized organizations, almost all of exceptions are all ways up to the top of

a hierarchy to ask his or her decision. In the case of the Japanese organization style, high

level of centralization and multiple levels of hierarchy persuade to ask exceptions to a

project manager. The project manager might cause large backlogs to handle with many

exceptions raised from his or her subordinates.

100

To summarize the components of hidden work, Japanese actors are effective in

reducing rework volume, and American actors are effective in reducing coordination and

wait-time volume. When considering the proportion of the three components of hidden

work, rework occupied the largest portion, about 75% of the total hidden work volume.

Thus, the greater the improvement in rework volume, the more effective the team

performance is. In the Japanese case, there is a better tradeoff between rework volume vs.

coordination and wait time. This tradeoff between rework volume vs. coordination and

wait time is derived from both micro-level behavior patterns and organization styles as

expected.

5.4 Discussion and Conclusion In this chapter, I have described an intellective experiment that examines the impacts of

changes in cultural values and cultural practices on project performance. The simulated

results are qualitatively compared to the two hypotheses: contingency theory (Thompson,

1967; Galbraith, 1974; 1977) and cultural contingency theory (Hofstede, 1991). This

section discusses four following respects: implications, limitations, validity and

conclusion.

5.4.1 Implications

The intellective experiment set the four basic types of workflow: a pooled, a sequential, a

reciprocal, and an intensive workflow. These four types of workflow represent the level

of project complexity from low to high (Table 5.10). In the experiment, three settings —

information exchange ratio, project error probability, and functional error probability —

that determine project complexity are shifted from low to high, corresponding to the

types of workflow (Table 5.11). These idealized project complexities are set in order

qualitatively to confirm the Galbraith’s hypothesis (Galbraith, 1974, pp28): “the greater

the uncertainty of the task, the greater the amount of information that has to be processed

between decision makers during the execution of the task in order to achieve a given level

of performance.” I assume that hidden work volume represents the amount of

information with which a project team needs to handle. Figure 15 and 16 clearly show

that the hidden work volume is increased as project complexity goes up from the pooled

101

to the intensive workflows. Therefore, the emergent simulated results confirm the internal

validity of the VDT simulation since the VDT model is developed as the extension of

information processing abstractions.

The simulated results are compared with contingency theory (Thompson, 1967;

Galbraith, 1974; 1977). Thompson (1967) proposed that successful organization style is

contingent upon project complexity. The simulated results confirm his hypothesis.

Figure 5.14 illustrates that: the A organization style show better performance in the cases

of medium project complexity than that of the Japanese organization style; and the

Japanese organization style shows better performance in the case of high project

complexity than that of the American organization style (Figure 5.14). Thus, successful

organization style — either Japanese or American organization style — is contingent

upon project complexity in terms of hidden work volume. The simulated results also

imply that hidden work volume is one measurement of “successful” organization styles.

Specifically, the Japanese organization style consistently shows better performance in

project quality risk than that of the American organization style (Figure 5.18). Japanese

people may emphasize or give priority to project integration and project quality respects

as project criteria. In other words, meaning of “success” would depend upon project

criteria. Therefore, I extend the Thompson’s hypothesis to which successful organization

style is contingent upon project complexity and project criteria.

The simulated results are compared with Hofstede’s proposition of “cultural

contingency theory” (Hofstede, 1991, p.152; Adler, 1997). Hofstede introduced

Mintzberg’s five coordination mechanisms and projected them onto a power distance-

uncertainty avoidance plane, giving examples of typical countries. Hofstede’s work

implies that each nation shows better performance if they use their own preferred

coordination mechanism. Each nation’s preferred style is also predicted from a power-

distance-uncertainty avoidance matrix (Figure 5.19).

102

1. Adhocracy 2. Mutual adjustment 3. Support staff GREAT BRITAIN

1. Simple structure 2. Direct supervision 3. Strategic apex

CHINA

1. Divisionalized form 2. Standardization of outputs 3. Middle line

USA

GERMANY 1. Adhocracy 2. Mutual adjustment 3. Support staff

FRANCE (Japan)

1. Full bureaucracy 2. Standardization of work process 3. Techno-structure

Low

Uncertainty A

voidance

High

High Low Power Distance Figure 5.19: Preferred Coordination Mechanism (Hofstede, 1991, p.152)

Note: This figure illustrates the typical organization structure predicted by “power distance index” and “uncertainty avoidance index.” “Power distance index” refers to the extent to which the less powerful members of organizations and institutions accept and expect that power is distributed unequally. “Uncertainty avoidance index” indicates the extent to which a culture programs its members to feel comfortable in unstructured situations such as unknown, surprising, and different from the usual. Uncertainty-avoiding cultures try to minimize the possibility of such situations by using strict laws and rules, and safety and security measures

The simulated results shown in Figure 5.15 and 5.16 indicate that the Japanese

micro-level behavior pattern is effective in reducing the total hidden work volume, when

using the Japanese organization structure, except for the case of pooled workflow.

Similarly, the American behavior pattern is positively correlated to the American

organization structure in cases of medium to high complexity. Thus, the results show a

correlation between the micro-level behavior pattern and the preferred organization

structure, which matches Hofstede’s theory, but in the limited condition of medium to

high task complexity.

When considering the components of hidden work volume for pooled workflow,

there is less rework volume than for other workflow types. This implies that the choice

of which micro-level behavior pattern leads to better performance depends on

103

characteristics of workflow. Specifically, if a workflow requires coordination efforts, the

Ameircan behavior pattern is more appropriate than the Japanese pattern. On the other

hand, if the workflow requires much rework, the Japanese pattern is more suitable.

Another implication of Hofstede’s theory is that the preferred coordination

mechanism can be predicted from a power distance-uncertainty avoidance matrix. Based

on case studies, the Japanese and American organization structures are close to the

preferred mechanism plotted by Hofstede. Specifically, the Japanese organization

structure has relatively high centralization, high formalization, and a pyramid type

configuration. Hofstede also suggests that Japan is categorized with France as a type

with a full bureaucracy (pyramid model). It was appropriate to set every parameter of the

American organization at the medium level, because America is located in the center of

the diagram (Figure 5.19). This may be one reason why the divisionalized form,

developed in the United States, enjoys such great popularity globally (Hofstede, 1991).

In the point of departure section (Chapter 2), I discussed why a specific

organizational form has been fostered and elaborated in a country as a part of the

isomorphism theory. Based on my literature survey, there are three possible prototypical

views: the culturalist view, the universalist view, and the istitutionalist view. What kinds

of implications can my simulated results contribute to this argument? As the first point,

the Japanese organization style consistently shows better project quality risk, matching

the high UAI value of Japan. This implies that the Japanese cultural value system may

have led to the Japanese organization style as an acceptable practice, thus confirming the

culturalist view. The second point is that each cultural team member type shows better

performance when using its own organization structure. This indicates that people in

each country may select or evolve the most efficient organization design based on their

distinctive behavior pattern. People may understand how they behave or react in their

home country’s environments and then accumulate a sense what is the most effective

organization form over years or decades, implicitly and/or explicitly. This finding

somewhat reinforces both the culturalist and universalist views, since individual behavior

patterns are largely influenced by cultural value systems (the culturalist view), and since

individuals make rational decisions on selecting an organization form (the universalist

view). Thirdly, my simulated results showed that the level of project complexity has a

104

bigger impact on team performance than changes in organization structures and

individual behavior patterns. This implies that an organization design needs to be

considered based not only on combinations of preferred organization structure and

individual behavior patterns (culturally-driven normative systems), but also given project

contextual variables — project complexity, uncertainty, and ambiguity. Therefore, my

simulated results can support the institutionalist view combined with the culturalist and

the universalist views.

5.4.2 Validity and Limitations

The main purpose of the intellective experimentation step was to assess encoded

behavioral and organizational parameter settings appropriately in the model. Based on

the implication sections, the effects of changes in micro-level behavior patterns and in

organizational control styles shows interesting correlations between values and practices,

and also affords evidence that these parameters are encoded, more or less, correctly by

comparing model predictions qualitatively to the extant theory such as organizational

contingency theory (Galbraith, 1974; Thompson, 1967) and cultural contingency theory

(Hofstede, 1991; Adler, 1997) with the limited conditions.

The existing VDT model has known limitations that constrained us in capturing

all of the cultural and broader institutional phenomena that emerge in global projects. I

was unable to adequately represent factors such as multiple behavior patterns for different

workers in a project, additional exceptions caused by work practice differences,

organizational learning, and some of the positive impacts — i.e., increased innovation —

that might result from cross-cultural interactions. Our experiment focused only on the

impact of different patterns of micro-level behaviors and organization structures.

- Examining the cases where multiple behavior patterns coexist in a project

remains an intriguing research focus. In the following section, I extend VDT

to permit a modeler to assign different cultural values to each “Actor” — i.e.,

each individual or sub-team — in the project

- A second constraint was that the current VDT model is not able to

parameterize additional exceptions caused by differing values and practices

between subgroups of a joint-venture team. In particular, based on our

105

observations, subgroups are likely to have their own standardized low-level

work practices, rules and criteria. Our ethnographies provided evidence that

such differences generated exceptions between subgroups when selecting

standardized criteria for a project, such as those used for safety. Several

researchers have addressed differences in institutionalized practices in IJV

projects (e.g., Mahalingam et al, 2004).

These limitations represent needs to extend the current VDT model to which a

simulation model can examine mixed cultural cases and incorporate additional exceptions

caused by differing values and practices (Chapter 6).

Additionally, this research began by calibrating encoded parameters in a

qualitative manner — i.e., 20% differences as the largest score gaps relative to the

American cultural value scores for making qualitative differences in individual behaviors.

Even though qualitative differences in individual behavior parameters showed interesting

results and stimulated arguments that were discussed in the implication section,

quantitative research remains high on the future research agenda.

5.4.3 Conclusion

In this chapter, I have described an intellective experiment that compared simulated

results with theoretical outputs in order to understand and analyze the effects of

culturally-driven normative systems on project performance. Culturally-driven

normative systems have two main constructs: cultural values and cultural practices.

Based on my ethnographies and the literature survey, the cultural values and practices are

encoded as micro-level behavior patterns and organization structure styles respectively.

In addition, the experiment addresses different project complexities and team situations

as idealized project contexts. These four inputs — micro-level behavior, organization

structure, project complexity, and team situation — are treated as independent variables

to varied through their full ranges to explore their separate and combined effects on

project performance (Figure 5.11). The simulated results are compared to the

hypotheses derived from: organizational contingency theory (Thompson, 1967; Galbraith,

1974; 1977) and cultural contingency theory (Hofstde, 1991).

106

First of all, the simulated results indicate that culturally-driven normative systems

do affect, by and large, project performance from an information processing point of

view. The impacts of cultural differences can be large as a project becomes coordination-

intensive and complex.

Secondly, the cultural values and practices comprising culturally-driven

normative systems have different impacts on team performance. For instance, changes in

behavior patterns had less impact on team performance than changes in organization

structure. At this stage, the relative contributions of the organization system or behavior

pattern are unknown and cannot be analyzed quantitatively.

The third finding is that each organization style has pros and cons. The American

organization style shows better performance in the cases of medium project complexity

than the Japanese organization style. However, in the case of high project complexity,

the Japanese organization style shows better performance than the American organization

style. Moreover, the American organization style is more sensitive to low team

experience than the Japanese organization structure.

Finally, one’s micro-level behavior pattern is positively correlated to one’s

organization style, for medium-high task complexity. This tendency is confirmed by

Hofstede’s proposition, “cultural contingency theory.” Understanding the relationships

between cultural practices and values through virtual computational experiments may

clarify the evolutionary phenomena of specific organization structures in each country

(Greif, 1994).

These conclusions provide initial evidence that cultural values and practices were

appropriately encoded in this simulation model. Using the current VDT model brings

limitations such as testing mixed cultural cases and modeling additional exceptions

caused by differences in practices.

107

108

CHAPTER SIX: PROTOTYPE MODEL,

INTERCULTURAL-VIRTUAL DESIGN TEAM

VDT research was initiated in the late 1980s with the long-term goal of developing new

theories and tools that could extend the reach of contingency theory and network-based

project management tools to analyze the performance of project organizations engaged in

complex tasks (Jin and Levitt, 1996). However, the trend toward globalization has

highlighted the coordination problems that can arise among people and groups in

international joint-venture projects. Therefore, the VDT research group turned toward

incorporating cultural and institutional factors into the modeling. The previous chapters

discussed about: what are the distinguished cultural and institutional factors in IJV

projects; and how to capture these factors using the information processing abstraction in

the VDT model through intellective experiments. My ethnographies and the first

intellective experiment found key limitations of VDT in attempting to analyze mixed-

cultural team cases. The main purpose of this chapter is to describe the prototype model I

developed to seek better organization designs for mixed-cultural teams. The model relies

upon observations based on my case studies to move beyond the “mono-cultural”

limitations of the current VDT model.

6.1 Purposes and Agenda of the IC-VDT Model The VDT model is described in (Jin and Levitt, 1996) and reviewed in Chapter 5.5. To

address the limitations of the current VDT in modeling culturally-driven organization

styles and decision making behaviors, this research develops a prototype computational

model with key extensions. The current VDT model can only represent technically

complex, but “mono-cultural,” engineering teams composed of actors who have a single

set of cultural values, belief, and norms. The first intellective experiment conducted in

Chapter 5 provides a proof of concept validation in using the VDT model for cross-

cultural (comparative) studies. The cultural value-and-practice dimensions were the key

concepts to incorporate cultural factors in the VDT model. The main purpose of this

chapter is to extend the current VDT model to represent and reason about organizational

109

performance for mixed-cultural teams, thus moving it toward intercultural studies. The

limitations that the first intellective experiment in Chapter 5 found are summarized as

follows:

- The current VDT model cannot examine cases where multiple behavior

patterns coexist in a project — e.g., An American project manager working

with Japanese engineers. Therefore, the prototype model needs to have a

function to assign multiple cultural values within a project team.

- The current VDT model is not able to parameterize differing cultural practices

between subgroups of a joint-venture team. In particular, based on our

observations, subgroups are likely to have their own standardized low-level

work practices, rules and criteria. Our ethnographies provided evidence that

such differences generated exceptions, named institutional exceptions,

between subgroups when selecting standardized criteria for a project, such as

those used for safety. Therefore, the prototype model needs to consider

institutional exceptions that arise from mixed cultural practices.

The prototype model, called “Intercultural Virtual Design Team (IC-VDT)”

model, aims to capture the coexistence of two different types of behavior patterns and

work practices within a team. IC-VDT is initially validated here using two cultural cases

— US and Japan — enabling simple experiments18 that are still complicated enough to

represent the case of mixed-cultural teams. Value differences and practice differences in

a team represent a kind of internal complexity for a global project team.

6.2 Framework for the IC-VDT Model Figure 6.1 illustrates the conceptual framework of the model, “Intercultural-Virtual

Design Team (IC-VDT).” IC-VDT is constructed by adding new features into the current

VDT model (Figure 6.1): national cultural value indices to represent input variables and

cultural mechanisms as elements of the simulation’s reasoning.

The current VDT model has four input variables: organization structure,

communication tools, project team descriptions, and project descriptions (see Figure 5.1).

This research adds national cultural indices in the project team descriptions. The modeler 18 The experiments here should be viewed as a prototype for examining multi-cultural cases.

110

can specify which national culture is involved in a project. The other inputs such as the

organization structure19, communication tools, and project descriptions are the same as

those in the current VDT model.

OUTPUTS INPUTS

Figure 6.1: Framework of the IC-VDT Model

The simulation system simulates the balance between information processing

demands and capacities. Information processing demands are determined by project

descriptions including task descriptions and uncertainty. Task descriptions specify direct

work volume, while uncertainty causes a certain rate of exceptions that contribute to

indirect work volume such as rework, coordination efforts, and wait time for decisions,

called hidden work volume (Levitt and Kunz, 2002). Information processing capacities

are determined by organizational structures and project team descriptions that include

19 When a project is assembled, a project manager determines his/her preferred organization structure. Thus, the global project managers have a relatively large amount of control over organizational practices. The current VDT has rich organizational parameters to represent organizational control practices. IC-VDT uses the same organizational parameters of VDT.

Organizational descriptions: Cultural Mechanisms Organization structure Project

Performance - Micro-level behaviors - Centralization, formalization, and organization configuration - Task control practices (Dependent

variables) Communication tools

Project contextual descriptions: VDT Project durationSimulationProject team descriptions - National Cultural Value Indices Project cost - Actors’ skill, experience, reporting

relationships, etc Technical Mechanisms - i.e., technically-driven exceptions, skill match

Project quality Project descriptions - i.e., interdependencies, task

Note: This figure illustrates adding new features into the original VDT model: national cultural value indices as input variables and cultural mechanisms in the simulation system. As input variables, a user can input national cultural value indices with scores ranging from 0 to 100. The IC-VDT simulation system is composed of two dimensions: cultural and technical dimensions. The cultural dimension represents the intercultural, internal complexity of global project teams, while the technical dimension represents task-driven information flow.

111

team experience and team members’ skills and application experience. I added the

cultural mechanism to the VDT model. The cultural mechanism (colored blue in Figure

6.1) is a new feature of IC-VDT and is composed of micro-level behaviors and task

control practices. Micro-level behaviors are linked to national cultural indices, and affect

the ways that information is processed in an organization (information processing

capacity). Task control practices are also linked to national cultural indices. They

determine the initial probability of institutional exceptions arising between cultural

groups, increasing the demand for information processing. The technical dimension

determines task-driven exceptions such as project exceptions20 and functional

exceptions21. Therefore, three types of exceptions — project, functional, and

institutional exceptions — affect the indirect work volume in IC-VDT.

As with the VDT model, IC-VDT predicts the simulated project duration, hidden

work volume, cost, functional risk and project quality risk. These dependent variables of

a simulation are the same as the current VDT model (Compare Figure 5.1 and Figure 6.1).

6.2.1 Value Differences

Cultural values refer to desirable criteria or standards for evaluating behaviors that people

show in making task-related and communication-related decisions. This research views

cultural value indices as the basis for the behavior patterns of individual actors in

decision-making and communication. These national cultural value differences are

implemented in two steps. First, a pre-processing function is created that generates two

sets of cultural behavior patterns by linking a baseline set of behavior parameters to a

national culture index specified by the modeler. For instance, in the case of American

culture, a modeler can input scores of PDI, IND, MAS, and UAI22 indices based on the

Hofstede’s research. The modeler also can test different set of value scores from the

original Hofstede’s research in order to understand changing value systems23 in a country

20 Project exceptions are work process-related errors that cause integration problems. 21 Functional exceptions refer to functional errors that are localized to a task requirement, causing defects of a product. 22 Power Distance Index (PDI), Individualism vs. Collectivism (IND), Masculinity vs. Feminity (MAS), and Uncertainty Avoidance Index (UAI) 23 Lachman (1994), for instance, addressed about the stability of cultural value systems. Lachman argued that core values in a country tend to be more stable and resistant to change than periphery values. However, we cannot deny a possibility to change the core values over years or generations.

112

and so forth. Second, IC-VDT allows the modeler to specify which of the behavior types

(in this case, “J” or “A”) each individual actor should possess. These two steps allow the

modeling and simulation of real “mixed-cultural team” cases.

Pre-Processor Model The IC-VDT model

Multiple sets of micro-

One set of micro-level behavior files (mono-cultural)

Inputs UAI PDI IND

User

National cultural indices

VDT Simulation Engine

f

Figure 6.2: Pre-Processor Model

Note: Users input national cultural value inautomatically generates multiple micro-levemodifying the original micro-level behavior

Pre-processor Model: The pre-proce

level behavior patterns by modifying the exis

values for the national culture index (Figure

as in IC-VDT, these behavior patterns are en

probability of choosing particular actions (Ch

(Hofstede, 1991) are used for national cultura

function enables the examination of other cu

micro-level behaviors is described in Chapte

assessed and validated encoded behavior par

The pre-processing model is impleme

can input each cultural score with a range bet

automatically calculates and creates two sets

six behavior parameters: 1) decision making

or ignore), 3) tolerance in waiting for decisio

response volume of communication, and 6) d

5.2.2). In my intellective experiment case, on

American behavior patterns, while another is

level behavior files

dices and then a pre-processor l behavior files by linking to and files.

ssing function generates two cultural micro-

ting behavior patterns when a user inputs

6.2) of the second, non-US.-culture. In VDT

coded in tabular files showing the

apter 5). Hofstede’s cultural dimensions

l indices. Changing the input values to this

ltural cases and sensitivity analyses. A set of

r 5.2.2. The first intellective experiment

ameters (Chapter 5.4).

nted using a MS-Excel spread sheet. A user

ween 0 and 100. The pre-processing model

of micro-level behaviors composed of the

policy, 2) type of decision (rework, correct,

ns, 4) attendance to communication, 5)

emand volume for communication (Chapter

e set of behaviors represents typical

for typical Japanese behavior patterns.

113

Figure 6.3: A Screenshot of Pre-Processing Model

Note: Users input national cultural value indices in input columns and then a pre-processor automatically generates multiple sets of micro-level behavior files by linking to and modifying the original micro-level behavior files. The right graph represents cultural group 1 and the left graph is for cultural group 2. For example, I input American cultural value indices on the left hand side, and Japanese cultural value indices on the right hand side. I used the American and Japanese scores provided by Hofstede (1991).

Figure 6.4: An Example of Micro-Level Behavior Parameters

Note: This figure illustrates the example of micro-level behavior parameters, decision making policy and time to wait for decision. Numbers in columns are linked to cultural value indices inputted and adjusted automatically. For more detailed calculation, please see Chapter 5.2.2. Appendixes show the two sets of micro-level behavior parameters, representing the typical A behavior and the typical J behavior patterns respectively.

114

Selection function: A user can select either J or A cultural micro-level behavior

patterns for each actor in IC-VDT. This function enables users to set a mixed cultural

case. This research preliminarily focuses on Japanese and American cultures. However,

the additional functions enable future researchers to examine and assess other cultural

cases such as China, Canada, and so on by simply entering appropriate values for the

Hofstede’s value dimensions. Figure 6.5 demonstrates that culture is an attribute of an

actor. A user can specify actor’s culture either Generic, Japanese or American. Generic

means the default behaviors, assumed to be American behaviors in this research.

Figure 6.5: A Screenshot of Selection Function

Note: This figure illustrates an example to select the culture of each actor from Generic, Japanese, or American cultures. Generic culture means the default behavior patterns of both VDT and IC-VDT that is assumed to be the American behavior pattern. Japanese culture is added in IC-VDT based on the literature survey, the case studies, and the observations. This selection function enables users to design multiple behavior patterns in a project. Therefore, they can design mixed cultural cases.

6.2.2 Practice Differences

IJV teams are composed of two or more cultural groups that are headquartered in

different countries. Each subgroup is likely to have its own typical, national set of

practices related to managing organizations and tasks. Practice differences refer to the

specific organization design used for coordinating organizations and tasks, adopted

according to the organizations’ national cultural norms. I observed in case studies that

work groups in IJVs spent a great deal of time in arguing about and coordinating

standards, rules and criteria for tasks. In the case of the Catwalk bridge project (Chapter

4), for instance, both the Japanese prefabricator and the American design firm insisted on

115

their own standardized ways of controlling quality and fabricating the steel products,

triggering a year long discussion between the parties. Similar phenomena are confirmed

by the preliminary findings of other CRGP research24. For instance, standardized rules

and criteria for construction safety used by Taiwanese contractors are very different from

those of Japanese contractors (Mahalingam, 2004; 2005). Empirical evidence shows that

each cultural work group is likely to have their own standard rules and criteria for tasks.

These rules and criteria are fostered and elaborated upon by institutional elements such as

the regulative, normative and cognitive systems in a country. In other words, the

standard rules and criteria have been institutionalized over years or decades. The

selection process engenders exceptions between the two subgroups, here called

“institutional exceptions.” Institutional exceptions are distinguished from technically-

oriented exceptions25. One of the distinguished characteristics of institutional exception

is that there is no absolute criterion, solution and process to solve this exception, since it

is not a matter of good or bad, rather a matter of preferences based on values. The

preliminary goal of the IC-VDT model is to address and model the concept of

institutional exceptions as a key cultural feature observed in IJV projects.

1. Institutional Exception-Generating Mechanism: Based on

empirical observations, IC-VDT must address differing institutional practices.

Figure 6.6 shows an abstract model of the institutional exceptions mechanism. A

team member (ST) is assigned to a certain task. Once a project exception is

generated, he or she notifies his/her boss (PM) to make a decision about the

project exception. When the ST receives the PM’s decision or when a PM

receives the notification message, the two actors may have different standards and

criteria for performing tasks if they are of different nationalities, generating an

institutional exception. The institutional exception causes the messages to

bounce between supervisor and subordinates, in turn causing delays in the time to

solve the institutional exception. However, if both PM and ST are from the same 24 Ashwin Mahalingam, a CRGP research colleague, has conducted ethnographic interviews in two large IJV projects in Taiwan and India (Mahalingam, 2004; 2005). Ryan Orr, also a CRGP research colleague, has conducted telephone interviews regarding cultural issues that arise in global projects (Orr, 2004; 2005). Their findings provided additional empirical evidence and insight. 25 Technically-oriented exceptions refer to exceptions which are generated by a mismatch between the required skill level of a task and the skill level of the responsible actor. High task uncertainty and complexity generally increase the frequency at which exceptions are generated.

116

cultural group, the institutional exception is less likely to occur. Moreover, the

time required to solve an institutional exception depends on the actors’ cross-

cultural experience and project-level team experience.

PM

(3) Make a decision (5) Re-notify an institutional exception

(2) Notify a message

(4) Generate an institutional exception

ST

(1) Generate a project exception Task

Figure 6.6: Institutional Exceptions Generating Mechanism

Note: This figure shows an abstract model of the institutional exception generating mechanism, based on case study observations. A “ST” is a team member, representing culture “Green”, and a “PM” is a project manager, from culture “Blue.” The project exception is the trigger for an institutional exception between the twcultures, “Green” and “Blue.” Notification messages go back and forth between the two actors, in an “exception bouncing phenomenon,” extending the time required tosolve the institutional ex

o

ception.

1. Determination of Institutional Parameters: With respect to

institutional exceptions, IC-VDT adds three new parameters: Institutional

Exception Probability, Institutional Exceptions Distribution Policy, and Cross-

Cultural Experience Effect on Priority.

(a) Institutional Exception Probability: when an actor reports to a supervisor

with a different cultural background, the two actors encounter differences in task

processes, goals, criteria, and interests, creating an "institutional exception." The impact

of "institutional exceptions" is calculated based on the following formula:

117

Total Volume of Institutional Exceptions = Global Weight (GE) x Total

Volume of Project Exceptions (PE) x Institutional Exceptions Probability (IEP)

Global Weight (GE) determines the initial volume of Institutional Exceptions

(IP) calculated as the proportion of the probability of project exceptions (PE) in a task.

IC-VDT views project exceptions as a trigger for exposing differences between each

actor’s preferences about processes, criteria, and objectives. Global Weight (GE) is

adjusted by the level of team experience. If a project team has experience working

together in previous projects, they are more likely to understand differences between

subgroups and to have experience in resolving issues related to differing institutions. On

the other hand, in the case of no previous team experience, they are less likely to

understand and are less able to resolve differences between subgroups.

Table 6.1: Global Weight Adjusted by Team Experience H M L

0.5 1 2

Note: This table shows the adjustment factor of Global Weight by Team Experience. Global Weight (GE) = 1 * Adjustment factor by Team Experience GE=1: means that each time a project exception occurs, there is (GE) probability of having an institutional exception. GE>1: means that conditional probability of institutional exceptions is greater than that of project exceptions. GE<1: means that conditional probability of institutional exceptions is less than that of project exceptions

Total Volume of Project Exceptions (PE) is determined by the global project

error probability and local task complexity, consistent with those determined by the VDT

model (Christiansen, 1993).

Institutional Exceptions Probability (IEP) determines the probability of

generating institutional exceptions based on the interface between actors. IC-VDT

assumes that the current institutionalized practices — i.e., quality, work process criteria,

and objectives — are embedded in their cultural contexts. Misperceptions between

participants can potentially take place in any of the five cultural categories: PDI, IND,

118

MAS, LTO, and UAI. Therefore, Institutional Exception Probability (IEP) is calculated

as follows:

Global Probability of Institutional Exceptions = (ΔPDI + ΔUAI + Δ

IND + ΔMAS + ΔLTO ) / 500 (%)

(b) Institutional Exceptions Distribution Policy (IEDP): This behavior,

encapsulated by a parameter matrix, is used by IC-VDT to determine the actor that

should make the decision for an institutional exception, based on the project's

centralization level. The general assumption is that more centralized project teams

require higher level participants to make decisions about resolving exceptions. This

matrix is the same as the Decision Maker Policy in the VDT model. This work assumes

the same distribution policy as VDT, since there is no empirical evidence available

through case studies about how the resolution of institutional exceptions might differ

across cultures.

Table 6.2: Institutional Exceptions Distribution Policy (IEDP) High Medium Low

PM 0.6 0.2 0.1

SL 0.3 0.6 0.3

ST 0.1 0.2 0.6

PMSL

ST

High

Medium

Low0

0.2

0.4

0.6High

Medium

Low

Figure 6.7: Institutional Exceptions Distribution Policy (IEDP)

119

Note: Table 6.2 and Figure 6.7 show the probabilities of the actor role that will make a decision about an institutional exception. For instance, in the case of medium level of centralization, a PM has a 20% probability of making a decision about an institutional exception. Under a highly centralized organization structure, this probability goes up to 60% (left front bar for high centralization and PM).

(c) Cross-Cultural Experience Effect on Priority: Individuals who have

worked for global projects in the past can certainly provide useful knowledge about how

to handle similar internal institutional exceptions. For instance, experienced individuals

in global projects know what types of problems were raised, their solutions, and the

probable results. These experiences provide useful information and knowledge for future

situations. On the other hand, a less experienced individual may not understand the types

of problems that can occur, how to handle these problems, and the consequences of

actions he or she selects. In our case studies, we observed many freelance expatriates

who contributed by providing this type of knowledge (Mahalingam, 2004). In addition to

this, Hofstede (1991, pp.232) and Adler (1997) pointed out that experienced freelance

expatriates can provide knowledge and skills that are applicable for any foreign cultural

environment. In IC-VDT, the level of each participant's cross-cultural experience

determines the level of priority he/she assigns to attending to institutional exceptions. In

general, the greater the amount of cross-cultural experience a participant has, the greater

priority the participant assigns to attending to institutional exceptions. On the other hand,

a participant who does not have cross-cultural experience is less likely to be aware of, or

to attend to, institutional exceptions. Therefore, the highly experienced participants

generally assign high priority to institutional exceptions, while less experienced

participants assign low priority. If an institutional exception has high priority, a decision

maker is more likely to pick it up quickly, processing and defusing the exception more

rapidly.

Table 6.3: Cross-Cultural Experience Effect on Priority High Medium Low

Priority level H M L

Note: This table shows the level of priority depending on the receiver’s cross-cultural experience. If an actor has high cross-cultural experience, he/she pays more attention to institutional exceptions. On the other hand, an actor with low cross-cultural experience is less likely to pay attention to institutional exceptions.

120

6.3 Contingency Fit Information processing theory (Galbraith, 1974, 1977; Thompson, 1967) underlies the

viewpoint of the IC-VDT model in construing organizations as information processing

systems. The interaction between information demand and capacity generates an

overflow or underflow of information, creating hidden work volume (Levitt and Kunz,

2002) and causing schedule delay and quality risks in cases of information overload.

This information processing “tipping point”, was discussed by a previous Stanford

researcher (Fyall, 2002) as the boundary between “laminar” vs. “turbulent” information

flow in an organization, analogous to the “Reynolds Number,” which predicts laminar vs.

turbulent fluid flow in pipes or open channels. The original VDT model is designed to

capture technically-driven information demand vs. capacity. The IC-VDT model adds

institutionally-driven information demand vs. capacity as an important cross-cultural

feature of global projects (Figure 6.8).

Information Processing Demand

Inin

P

C

IInternal institutional Complexity

Institutionally–driven information demand

Technically–driven information demand Uncertainty

Technically–driven information capacity

Actor capacity

Task Requirements

Task Complexity

Task Interdependency

Figure 6.8: Information Demand and Capacity

Note: Figure 6.8 illustrates the balance between the information dinformation processing capacities of the IC-VDT model. There atechnically-driven information balance and institutionally-driven Technically-driven information balance is simulated by the currenIC-VDT model addresses institutionally-driven information balan

Actor’s behavior pattern (Values)

Leadership style(Practices)

Information Processing Capacity

stitutionally–driven formation capacity

roject Team Experience

ross-cultural Experience

nformation Exchange

emands vs. the re two layers: information balance. t VDT model. The ce as a second layer.

121

Information demand: There are two categories of information demands:

technically-driven and institutionally-driven demands. Technically-driven information

demand is defined by uncertainty, task complexity, task skill requirements, and task work

volume (Jin and Levitt, 1996). Institutionally-driven information demand is created by

the differences in institutionalized practices between sub-groups. This research

investigates the different institutionalized practices (task control style) caused only by

national cultural norms. The impact of different professional norms remains an

interesting area for future research. Institutionally-driven information demands are

generated according to the degree of internal complexity of IJV teams. They represent

the difficulty of coordination among subgroups that come from countries with different

cultural norms and different task control systems.

Information capacity: In parallel with demand, there are two categories of

information processing capacity: technically-driven information processing capacity and

institutionally-driven information processing capacity. The technically-oriented

information capacity is determined by an agent’s capacity, skill set, application

experience, project leadership style, and behavior patterns (Jin and Levitt, 1996).

Institutionally-driven information capacity is determined by information exchange,

project leadership style, and experiences. These two attributes represent ways to solve

institutional exceptions. One way to handle institutional exceptions is to take a

decentralized leadership style. In the case of low centralization (decentralization), for

example, a project manager delegates authority to subordinates. Less interaction between

actors can substantially reduce institutional exceptions. The second option is to handle

exceptions using individual and/or group experience. An experienced agent can provide

the knowledge of how to handle a type of exception from his or her past experience.

Third country expatriates26 may play a role in exchanging this type of information and

knowledge by accumulating cross-cultural experience and knowledge as they move

between successive global projects. High team experience27 can also mitigate conflicts

26 In the SC project, we observed many expatriates working for international joint-venture projects. My colleague also observed expatriates in Asian countries who originally came from England and worked as contract employees for participating firms from multiple countries in international projects (Mahalingam, 2004). 27 High team experience means that subgroups composing the project team have past experience in working together.

122

regarding the selection of practices, because the groups are already experienced with

working together and have had substantial prior experience at resolving these kinds of

exceptions.

6.4 Summary of the IC-VDT Model 6.4.1 Overview

The values-practices dimensions address the internal complexity within a global project

team. At the individual level, each actor has different values, leading to different

behavior patterns of decision making and communication. IC-VDT starts by creating two

different types of micro-level behavior patterns: typical Japanese and typical American

patterns (upper-right hand side in Figure 6.9). Furthermore, a user can allocate two

different types of behavior patterns, distinguished by value differences, for each actor

who has a role such as project manager (PM), sub-team leader (SL), or sub-team member

(ST) in an organization chart (upper-left hand side in Figure 6.9). This function allows

users to examine and analyze mixed cultural cases.

On the practices dimension, organizational practices are related to the

organization structure such as the level of formalization, the level of centralization, and

the organization configuration (upper-left hand side). Managers have relatively good

control over organizational practices.

Project descriptions are composed of four elements: complexity, uncertainty,

interdependencies and requirements. These elements define the technically-driven

demand for information processing.

The center part represents an interface between the information demand and the

information capacity. When one actor is assigned to a certain task, one needs to process a

certain amount of information that is defined by the technically-driven demand for

information processing. The project descriptions — i.e., project uncertainty and/or skill

matches between task requirements and actor’s skills — define the likelihood that one

actor has a technically-driven exception. When one actor gets an exception, the actor

notifies a message to his/her super-ordinate. When two actors are of different

nationalities, they are most likely to have different task control styles — i.e., different

preferred standards, rules, and criteria. These differences engender exceptions, called

123

“institutional exceptions.” Therefore, IC-VDT sees a technically-driven exception as a

trigger to generate an institutional exception. The probability of generating institutional

exceptions depends on not only the interfaces between actors, but also the cultural value

gaps between two countries.

Figure 6.9: Framework of the IC-VDT Model

Project Descriptions

Pattern A

Pattern J

Decision making parametersCommunication parameters

Decision making parameters Communication parameters

Selections

Start

Successor

Finish

In-tray Out-tray In-tray Out-tray

Organization structure Practice (Organization Structure) Values (Micro-level behavior)

Info

rmat

ion

Cap

acity

PM - Centralization - Formalization - Matrix Strength SL - Configuration

ST Organization chart

Communications to other actors

Technically-driven demand for information processing

Inte

rfac

es

Notify a message Actor A Actor J

Generate an institutional exception

Info

rmat

ion

Dem

and

- Project complexity - Project uncertainty - Task interdependencies- Task requirements

Workflow chart

Note: This figure summarizes the conceptual framework of IC-VDT. The upper part of the figure represents the information capacity of a team

composed of practices and values. Values mainly operate at the individual level. Thus, IC-VDT starts by creating two different types of micro-level behavior patterns: typical Japanese and typical American patterns (the upper-right hand side). Practices refer to the organization structure that is comprised of the four elements: centralization, formalization, matrix strength, and organization configuration. A user can set an organization structure style to explore its performance. The user also can allocate two different types of behaviors, distinguished by value differences, for each actor who has a specific role such as PM, SL, or ST in an organization chart (the upper-left hand side).

124

The lower part represents information demands. Project descriptions are composed of four elements: complexity, uncertainty, interdependencies and requirements. These elements define the technically-driven demand for information processing.

The center part represents an interface between the information demand and the information capacity. When one actor is assigned to a certain task, one needs to process a certain amount of information that is defined by the technically-driven demand for information processing. When a technically-driven exception is generated based on uncertainty and/or skill matches, the actor notifies a message to his/her subordinate. When two actors are from different countries, they are most likely to rely on different task control styles — i.e., different standards, rules, and criteria. These differences engender exceptions, called “institutional exceptions.” Therefore, IC-VDT sees a technically-driven exception as a trigger to generate an institutional exception. The probability of generating institutional exceptions depends on the interface between a sub-ordinate and a super-ordinate and the cultural value gaps between two countries.

6.4.2 Implementation of IC-VDT

IC-VDT is implemented using the Python dynamic programming language, which offers

good flexibility in extending IC-VDT and adding additional features to support future

research needs. The platform consists of a model editor, simulation engine, and a

charting and reporting module. These components can be used in a stand-alone fashion,

or combined together into a complete desktop modeling and simulation application.

The model editor provides the primary user interface to the model. It allows for

the construction of complex project models by dragging, dropping and connecting simple

graphical objects representing actors, tasks, meetings, etc., along with the relationships

between these objects. The model editor also includes a feature to edit the properties of

each object.

IC-VDT incorporates a discrete event simulation engine, which is conceptually

similar to the VDT implementation, but significantly extends its capabilities. The

simulation algorithm is now completely time-scale independent, allowing for consistent

results from tasks ranging in time duration from minutes to years. Certain model

properties, such as skill and experience levels, have been switched from course-grained

discrete values, to continuous numeric ranges, allowing for support of gradual learning

processes (Ramsey and Levitt, 2005). To support several ongoing projects of our

research group, Ramsey has implemented several conceptual extensions required by

125

various research projects, including knowledge management, trust issues, and cross

cultural elements, and developing POW-ER (Ramsey and Levitt, 2005), an extensible

programming interface to the VDT simulation system.

Figure 6.10: A Screenshot of IC-VDT

Note: This figure is a screenshot of IC-VDT. Shown above is a simple test case. The diagram represents actors, tasks, milestones, and meetings, along with the relationships between them such as rework links (red lines), communication links (green lines), task assignments (blue lines) and successors (black lines)

A user can set up number of trials per a simulation. The default number is 100

trials per a simulation. Outcomes are a mean of 100 trials. Graphical modeling and

outputs allow any users easily to start using IC-VDT. IC-VDT can be used for research,

management analysis, and/or an educational and training tool.

126

CHAPTER SEVEN: INTELLECTIVE EXPERIMENTS

FOR MIXED-CULTURAL TEAMS

The main purpose of this chapter is to understand and analyze the effects of mixed-

cultural cases. IC-VDT allows us to assess and investigate how IC-VDT can be used to

find better organization designs for mixed cultural teams. Similarly to the mono-cultural

cases (Chapter 5), this chapter uses the intellective experiment approach not only to

understand the effects of mixed cultural cases, but also to examine which organizational

strategies make culturally diverse teams most effective. In particular, four relevant

hypotheses are tested through the intellective experiment — i.e., using a decentralized

strategy allows sub-teams to take multiple forms, leading to better performance (Adler,

1997). The followings are sub objectives of this chapter:

- Address consistency of IC-VDT results with the current VDT model results for

internal validity

- Compare the simulated results of IC-VDT with empirical findings from the

literature

- Understand the effects of mixed cultural cases

- Understand the effects of changes in organizational styles

- Understand the effects of changes in matrix strength

- Understand significance and strength of variables such as organizational

settings, individual behaviors, and task complexities (hypothesis testing)

7.1 Hypotheses for Managing Mixed-Cultural Teams Will a cross-cultural group be productive or unproductive? What organization structures

does a leader need to set up for mixed-cultural team members? Based on the literature

survey, this research examines four relevant hypotheses: contingency theory, divergence

vs. convergence styles, increased coordination costs, and high vs. low matrix strength.

Contingency theory: Contingency theory is the basis of both VDT and IC-VDT

models. In order to confirm whether IC-VDT is consistent with VDT, this intellective

127

experiment addresses the same project contexts in both modeling environments used in

the first intellective experiment (Chapter 5). This is called internal validity. In other

words, the first hypothesis is to make sure whether IC-VDT is in line with contingency

theory (Thompson, 1967; Galbraith, 1973; 1974).

• Hypothesis 1-a: The greater the project complexity, the greater the

information processing requirements the project team has to address

• Hypothesis 1-b: The successful organization style is contingent

upon project complexity

Task interdependencies such as pooled, sequential, reciprocal and intensive

workflows represent from low to high project complexities. Information processing

requirements can be seen in total hidden work volume as a simulated result of IC-VDT.

Divergence vs. convergence styles: Adler argues that there are two styles to

manage mixed-cultural teams: divergence vs. convergence styles. A divergence style is

related to a decentralized structure that allows participants to have different opinions and

to stimulate creativities by interactions. A convergence style refers to a centralized

structure with a hierarchical organization. Divergence vs. convergence can be linked to

American vs. Japanese organization styles, respectively. For instance, American firms

tend to employ a divergence organization style with delegated authority and

responsibility, decentralization, and a flat hierarchy. Divergence is important for

stimulating innovative ideas and allowing multiple objectives. Less formalization

encourages more creativity and innovation. On the other hand, the typical Japanese

organization style is to use high centralization and high formalization, with multiple

levels of hierarchy.

Adler suggests a “divergence” style during the early stages of project, because a

project team creates ways of defining its objectives, gathering and analyzing information

and developing alternative forms of action, and because a divergence style enhances

creativity and innovation (Adler, 1997). On the other hand, a convergence style

becomes important during the final stages of projects, since teams need to agree, or

128

converge, on which decisions and actions to take. In addition, Adler argues that cultural

diversity makes work processes easier during the earlier stages, because project teams

can employ a divergence style to take advantage of differences (Adler, 1991). On the

other hand, cultural diversity makes work processes difficult during the final stages, since

project teams need to converge and integrate all ideas and options. In this regard, global

projects may have more exceptions during the final stages, or at intermediate stages

requiring convergence.

This research combines the above issues by calling the American organization

style the divergence style and a model for the early stages of projects. It calls the

Japanese organization style the convergence style and a model for the final stages of

projects.

Hypothesis 2: The convergence style (Japanese) shows poorer performance than

the divergence style (American) in the early stages of projects and conversely.

The matrix swing concept from Morris (1982) argues that a project’s task

complexity in terms of interdependence between tasks — i.e., pooled, sequential, and

reciprocal workflows — changes systematically over the life cycle of a project — i.e.,

reciprocal workflow during the project shaping phase (Millar and Lessard, 2000),

sequential workflow during the implementation phase, and reciprocal and/or sequential

workflows during the turnover phase. He argues that organization styles needs to adapt

accordingly — i.e., a centralized structure for the project shaping and the turnover phases,

and a decentralized structure for the implementation phase. This life cycle pattern of

construction projects helps us understand which stage is more critical, and how to change

organization styles in accordance with a project in progress. The American organization

style that has a decentralized structure implies fitting into the implementation stage. On

the other hand, the Japanese organization style with a centralized structure is most likely

better at the beginning or at the end. Moreover, by overlapping the Morris’s matrix

swing concept (1982) to mixed cultural team cases, emergent simulated results may

interpret organizational behaviors of international joint ventures.

129

Increased coordination cost: Levitt argues that global projects provide an ideal

field setting in which to explore the effects of institutional clashes on the behavior and

performance outcomes of organizations, through increased coordination costs (Levitt et al,

2004). Thus, mixed cultural teams are likely to have increased coordination costs

compared to single cultural teams. Adler argued that culturally diverse teams often

perform below expectations and organizational norms when cultural differences are

ignored and poorly managed (Adler, 1983; Adler, 1997). My case studies also conclude

that mixed cultural teams perform less efficiently than expected in three out of the four

cases (Chapter 4).

Hypothesis 3: Mixed cultural teams have greater information processing demand

than single cultural teams

Matrix structure: One of the classic solutions to managing a mixed culture team

is the use of a matrix structure (Hofstede, 1991). The matrix structure refers to

organizations that employ a multiple authority system that includes not only multiple

authorities existing in a structure but also related support mechanisms and an associated

organizational culture (Davis and Lawrence, 1977). In the case of a mixed-cultural team,

there could be, for instance, two managers in a business unit, one who coordinates a

particular function across all regions, and another who coordinates all functional units in

the particular project. Both managers share their authority and status over common

subordinates. A multiple manager model can mitigate cultural differences since

subordinates are supervised by both managers, each representing a different cultural

background. Davis and Lawrence highlighted Citibank as a successful example of an

organization that uses a matrix structure to manage multiple cultures and international

markets (Davis and Lawrence, 1977). One key factor is to use nationally based

organizations and balance geographic and business market units under national

management. This is becoming a model for organizing and managing global

corporations in both the industrial and service sectors. The success of this model implies

that organizational connectedness (which corresponds to how close the workers are

geographically) is an important factor for a successful matrix structure in a global

130

corporation. A high degree of matrix strength can have better performance for

multicultural teams.

Hypothesis 4: The greater the strength of the matrix organization, the fewer

exceptions for cross-cultural teams

7.2 Parameters Used for Intellective Experiments The first intellective experiment (Chapter 5) evaluated encoded micro-level behavior

parameters determined through ethnographic observations, and provided interesting

correlations between values and practices. However, the current VDT model cannot

examine mixed cultural teams, limiting the usefulness of VDT in modeling international

projects. Chapter 6 modifies and develops a prototype model, IC-VDT, to model the

multiple behavior patterns of actors in a project and incorporate institutional exceptions.

In this chapter, I describe a second intellective experiment, where combinations of

multiple micro-level behavior patterns are tested in a proof-of-concept experiment.

Figure 7.1: Framework of Intellective Experiment

Note: This figure illustrates the framework of the intellective experiment for mixed cultural teams. There are five independent variables. Each variable has 1, 2, 3, or 4 possible cases. Thus, there are total of 96 possible scenarios. The outcome metrics, the dependent variables of the experiment, are duration, cost, and quality risks.

131

Figure 7.1 shows the design of the intellective experiment used to capture the

impact of mixed cultural teams on a construction project. The inputs of the intellective

experiment have two constructs: team descriptions and project descriptions. Team

descriptions have three attributes that affect an organization’s information processing

capacity: (1a) organization styles, (1b) matrix strength, and (2) combinations of micro-

level behaviors. The project description can be broken down into two elements: (3) task

complexity and (4) team situations.

(1a) Organization Style: This research sets up two different organization

styles: the American organization style and the Japanese organization style. The

American organization style has decentralized authority, a medium level of formalization,

and flat organizational hierarchy, representing a divergence structure. On the other hand,

the Japanese organization style has centralized authority, a high level of formalization,

and multiple levels of hierarchy, exemplifying a convergence structure. Consistent with

Adler’s hypothesis, divergence structures yields better performance for culturally diverse

teams (Adler, 1997). Adler also observes that all projects need to integrate or converge

at the final stages of a project. Therefore, each cultural organization style struggles or

faces problems at different stages of projects.

Table 7.1: American and Japanese Organization Styles

Leadership Style Type J Type A

Centralization High Low

Formalization High Med

Organizational Configuration Multiple layers of

hierarchy Flat level of hierarchy

(1b) Matrix strength: In IC-VDT, the matrix strength models the

"connectedness" of an organization by setting the probability that workers will attend to

exchanges of information. The three types of information exchange that IC-VDT models

are meetings, communications about tasks, and noise. Organization connectedness often

corresponds to how near the workers are geographically. This organization

132

connectedness is a key success factor of Citibank (Davis and Lawrence, 1977). For

example, if all workers are in one large room, communication will typically be informal

with little need for formal face-to-face or virtual meetings. If workers are distributed

across the country, there will be a greater need for meetings and fewer communications

in the hallways or across the room. Therefore, workers in an organization with high

matrix strength tend to engage in more informal communication, so high matrix strength

complements low formalization. Where there is low matrix strength, workers tend to have

more formal meetings, so low matrix strength complements high formalization. This

intellective experiment examines one variable of the hypothesis presented by previous

researchers (Hofstede, 1991; Davis and Lawrence, 1977), organizational connectedness.

It is a key success factor in managing a global team.

Table 7.2: Matrix Strength

Matrix Strength High Medium Low

(2) Combinations of micro-level behaviors: There are four possible

combinations of micro-level behaviors: All American team members (Type 1), American

team with a Japanese subgroup (Type 2), Japanese team with an American subgroup

(Type 3), and all Japanese team members (Type 4). Type 1 and 4 represent single

cultural teams, while Type 2 and 3 are mixed cultural teams. This chapter uses the same

organization structure as the intellective experiment described in Chapter 5. These

experiments use the structure of seven people in a team including one project manager,

two sub-team leaders, and four sub-team members. The actor’s hourly salary varies with

the roles of team members (PM, sub-team leader or sub-team member). Each individual

actor is assigned either the American or Japanese micro-level behavior pattern. The

micro-level behavior patterns used are explained in Chapters 4, 5, and 6. All actors have

exactly the same skill level, specialty, application experience, and the same capacity of

one FTE.

133

Table 7.3: Combination Patterns of Teaming

Type Type 1 Type 2 Type 3 Type 4

Single cultural

team

Mixed cultural

team

Mixed cultural

team

Single cultural

team

Project

Manager (PM)

Pattern A Pattern A Pattern J Pattern J

Subordinates

(SL, ST)

Pattern A Pattern A

Pattern J

Pattern A

Pattern J

Pattern J

# of AM to JP 7:0 4:3 3:4 0:7

(3) Project complexity: Like the previous intellective experiment, I set up

four levels of task interdependencies — pooled, sequential, reciprocal, and intensive

interdependencies (Thompson, 1967; Bells and Kozlowski, 2002). These four levels of

interdependency represent a range from low to high project complexity.

Table 7.4: Task Interdependency and Complexity

Pooled Sequential Reciprocal Intensive

Communication Error Prob. 0.4 0.47 0.53 0.6

Noise Error Prob. 0.1 0.1 0.1 0.1

Functional Error Prob. 0.1 0.1 0.1 0.1

Project Error Prob. 0.08 0.09 0.11 0.12

Institutional Error Prob. 1 1 1 1

Communication links 6 7 11 13

Rework links 0 5 10 15

CPM duration 200 days 1200 days 700 days 400 days

Total work volume 1400 days 1400 days 1400 days 1400 days

134

(4) Team experience: Because most international joint venture projects do not

have any previous experience working together, only the case of low team experience is

tested. In addition, the first intellective experiment showed that low team experience

cases magnify the impacts of cultural differences. This implies that low team experience

is a trigger to cause misunderstandings or conflicts between cultural groups in a real

situation. The secondary purpose of this intellective experiment is to investigate how

mixed-cultural teams can cope with low team experience.

7.3 Experimentation As shown in Figure 7.1, I simulated a total of 96 scenarios (3 matrix strength x 2

organization styles x 4 combination of micro-level behaviors x 4 task complexity levels x

1 team experience level). For experimental purposes, the actor and task configurations

are identical28. The IC-VDT model is designed to predict duration, cost, quality risks and

project risks as measures of team performance. The IC-VDT model shows the simulated

and the critical path method (CPM) duration. The simulated duration is calculated by

considering the simulated work volume and workflow. The CPM duration is calculated

by considering the designed work volume and workflow. The gap between the simulated

and the designed work volume is called “hidden work” (Levitt and Kunz, 2002), and is

caused by rework, coordination efforts, and wait time for decisions. Thus, this hidden

work is inversely correlated to the efficiency of team performance. The cultural practice

and values differences in the Japanese vs. American structures and the micro-behaviors

of the actors cause differences in hidden work volume, and hence in schedule, cost and

quality outcomes. Three dependent variables are analyzed, 1) hidden work volume, 2)

product quality risks29, and 3) project quality risks30, to analyze the impacts of changes in

organization styles and micro-level behaviors on team performance.

28 Actor and task configurations include actors’ skills, skills required by tasks, duration of tasks, hourly salary of actors, total number of team participants (all teams are composed of 7 members including 1 project manager, 2 sub team leaders, and 4 sub team members), and task responsibility assumptions. 29 Product quality risk index represents the likelihood that specialized components produced by this project will have defects based on failed rework and exception handling. 30 Project quality risk index represents the likelihood that the overall product produced by this project will not be integrated at the end of the project, or that the integration will have defects based on failed cross-disciplinary rework and exception handling.

135

7.3.1 Simulated Results

Tables 7.5-8 show the summary of the simulated results, including duration, hidden work

volume, cost, functional quality risks, and project quality risks.

Table 7.5: Simulated Results of Pooled Cases A Org. J Org.

Combination A A-J J-A J A A-J J-A J

CPM duration

Duration (M) 10.81 11.00 11.04 10.81 10.64 10.81 10.83 10.64

Standard dev 0.10 0.10 0.11 0.11 0.10 0.11 0.11 0.10

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Cumulative Hidden Work Volume

Duration (M) 51.21 51.14 51.08 51.25 50.89 50.88 50.96 51.02

Standard dev 0.31 0.29 0.30 0.34 0.34 0.28 0.32 0.32

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Cost

Cost ($1,000) 791.0 789.5 788.6 785.9 785.9 785.3 786.7 787.8

Standard dev 4.9 4.5 4.7 5.5 5.5 4.8 5.2 5.1

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Functional Quality Risk

Functional

Quality Index 0.40 0.39 0.39 0.38 0.40 0.39 0.38 0.37

Standard dev 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Project Quality Risks

Project Risk

Index (PRI) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Standard dev 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Two sample tests

(n=100) - - - -

136

Table 7.6: Simulated Results of Sequential Cases A Org. J Org.

Combination A A-J J-A J A A-J J-A J

CPM duration

Duration (M) 63.84 65.84 65.11 64.18 66.10 66.58 66.64 65.81

Standard dev 1.07 3.84 3.17 1.01 0.71 0.97 1.01 0.74

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Cumulative Hidden Work Volume

Duration (M) 56.66 57.63 57.24 56.75 57.53 57.09 57.12 57.51

Standard dev 1.24 3.11 3.04 1.16 0.76 0.80 0.83 0.77

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Cost

Cost ($1,000) 882.6 893.2 890.3 884.8 884.8 878.3 879.7 886.3

Standard dev 17.1 40.7 43.7 11.0 11.0 10.9 11.4 10.9

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Functional Quality Risk

Quality Index

(FRI) 0.40 0.39 0.38 0.37 0.40 0.39 0.39 0.37

Standard dev 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Project Quality Risks

Project Risk

Index (PRI) 0.52 0.53 0.54 0.41 0.33 0.41 0.49 0.31

Standard dev 0.06 0.08 0.10 0.06 0.05 0.05 0.04 0.04

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J > Type J

Type J-A > Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

137

Table 7.7: Simulated Results of Reciprocal Cases A Org. J Org.

Combination A A-J J-A J A A-J J-A J

CPM duration

Duration (M) 45.11 58.40 51.17 57.40 37.80 45.34 43.24 36.94

Standard dev 1.54 3.31 2.41 3.37 1.87 5.40 4.24 0.80

Two sample

tests (n=100) Type A < Type A-J

Type A < Type J-A

Type A-J < Type J

Type J-A < Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Cumulative Hidden Work Volume

Duration (M) 127.02 135.42 129.23 165.06 74.22 97.93 92.73 72.43

Standard dev 5.51 6.06 5.76 10.21 6.64 13.51 10.70 3.42

Two sample

tests (n=100) Type A < Type A-J

Type A = Type J-A

Type A-J < Type J

Type J-A < Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Cost

Cost ($1,000) 1918.5 2013.6 1943.7 1118.5 1118.5 1445.2 1375.8 1100.7

Standard dev 81.0 87.6 86.5 86.2 86.2 175.5 143.6 45.9

Two sample

tests (n=100) Type A < Type A-J

Type A = Type J-A

Type A-J > Type J

Type J-A > Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Functional Quality Risk

Quality Index

(FRI) 0.40 0.39 0.39 0.37 0.40 0.39 0.39 0.38

Standard dev 0.02 0.02 0.02 0.02 0.02 0.02 0.25 0.03

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Project Quality Risks

Project Risk

Index (PRI) 0.50 0.52 0.51 0.41 0.33 0.47 0.50 0.32

Standard dev 0.01 0.01 0.01 0.01 0.03 0.02 0.03 0.02

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J > Type J

Type J-A > Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

138

Table 7.8: Simulated Results of Intensive Cases A Org. J Org.

Combination A A-J J-A J A A-J J-A J

CPM duration

Duration (M) 65.57 77.37 74.94 94.50 33.97 38.27 38.04 31.84

Standard dev 4.30 4.70 4.54 7.24 5.17 5.84 6.17 2.34

Two sample

tests (n=100) Type A < Type A-J

Type A < Type J-A

Type A-J < Type J

Type J-A < Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Cumulative Hidden Work Volume

Duration (M) 178.74 176.75 179.04 273.66 87.50 91.05 89.46 81.52

Standard dev 11.49 10.03 10.92 20.34 16.42 18.25 16.78 6.39

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J < Type J

Type J-A < Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Cost

Cost ($1,000) 2628.7 2575.4 2621.2 1313.3 1313.3 1364.0 1332.8 1228.0

Standard dev 164.4 144.1 160.3 235.8 235.8 258.5 237.8 92.8

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J > Type J

Type J-A > Type J

Type A = Type A-J

Type A = Type J-A

Type A-J > Type J

Type J-A > Type J

Functional Quality Risk

Quality Index

(FRI) 0.40 0.39 0.38 0.37 0.40 0.39 0.38 0.37

Standard dev 0.02 0.02 0.02 0.02 0.03 0.03 0.02 0.02

Two sample

tests (n=100) Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Type A = Type A-J

Type A = Type J-A

Type A-J = Type J

Type J-A = Type J

Project Quality Risks

Project Risk

Index (PRI) 0.50 0.52 0.52 0.41 0.34 0.49 0.49 0.33

Standard dev 0.01 0.01 0.01 0.01 0.02 0.03 0.03 0.02

Two sample

tests (n=100) Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

Type A < Type A-J

Type A < Type J-A

Type A-J > Type J

Type J-A > Type J

(See notes on following page.)

139

Notes: (1) Total simulated work volume is the sum of production work volume and

coordination work volume (Jin and Levitt, 1996) Hidden Work Volume = Total Simulated Work Volume – Designed Work

Volume (2) For each scenario, 100 trials are run and means and standard deviations are

calculated. (3) Product quality risk represents the likelihood that components produced by the

project have defects based on rework and exception handling (Jin and Levitt 1996) (4) Project quality represents the likelihood that the components produced by the

project will not be integrated at the end of the project, or that the integration will have defects based on rework and exception handling (Jin and Levitt, 1996).

(5) All cases have matrix strength set to medium level

To fulfill the objectives of this chapter, I analyze simulated outcomes from three

points of view: (1) the impacts of changes in team combinations, (2) the impacts of

changes in matrix strength, and (3) the impacts of changes in organization styles. In

addition, I look into how institutional exceptions affect hidden work volume.

1. Single-cultural teams vs. Mixed-cultural teams: The intellective

experiment considers four possible combinations of behavior patterns. A and J represent

single cultural team cases, while A-J and J-A are cases of mixed cultural team. The A-J

team is composed of one American project manager, one American sub-team leader, one

Japanese sub-team leader, two American engineers and two Japanese engineers. On the

other hand, the J-A team is that the project manager position is replaced by Japanese.

other team members are same as the A-J team. The table 7.8 shows the average impacts

of changes in team member combinations, in particular from a single cultural team, A or J,

to a mixed cultural team, A-J or J-A. Overall, mixed cultural teams show greater project

duration (average 7% increased), greater cumulative hidden work volume (average 2%

increased), and greater project quality risk than single cultural teams (average 28.3%

increased). In particular, project quality risk is greatly increased in mixed cultural teams.

In a IC-VDT/VDT sense, project quality risks refer to the likelihood that the components

produced by the project will not be integrated at the end of the project. In other words,

IC-VDT results indicate that mixed cultural teams inherently have a large integration risk

140

rather than a functional or technical risk (See figure 7.4: Team combination patterns do

not affect product quality risk).

Table 7.9: Average of Increased Duration, Work Volume, or Quality

Average of increased duration, work volume, or quality

Project Duration 7.0% increased

Hidden Work Volume 2.0 % increased

Project Quality Risk 28.3 % increased

Note: This table illustrates the average impacts of changes from single cultural teams tomixed-cultural teams under the same project complexity. The outcomes are measured as the averages of increased project duration, work volume, or quality for all cases. X 100

Outcomes of mixed cultural teams utcomes of single cultural teams

*Average of all cases

O*% =

W

actually

and Figu

can see,

have gre

distinctio

of mixed

This incl

better pe

potential

organiza

hen taking a look at in detail, there is a case for which a mixed cultural team

shows better performance than the single cultural teams. For instance, Figure 7.2

re 7.3 show hidden work volume in the cases of reciprocal workflow. As you

in the case of Japanese organization style, the mixed cultural teams (A-J or J-A)

ater work volume than the single cultural cases (A or J). However, this

n is not clear in the case of the American organization style. The performance

cultural teams is similar between American and Japanese single cultural teams.

udes three implications: there is a potential that mixed cultural cases can show

rformance than the single cultural cases; the American organization style can

ly minimize the impacts of having mixed cultural teams; the Japanese

tion style potentially has less tolerance to have mixed cultural teams.

141

Hidden Work Volume (J organization style)

0

50

100

150

200

A J A-J J-A

Combinations of Team Members

A

J

A-J

J-A

Hidden Work Volume (A Organization style)

0

50

100

150

200

A J A-J J-A

Combinations of Team Members

A

J

A-J

J-A

Figure 7.2: Hidden Work Volume (J Organization Style): Case of ReciprocalWorkflow

Note: Figure 7.2 and 7.3 exemplifies the projmonth), for reciprocal cases. The X axis show

- A: American team: All team members hav- A-J: Mixed cultural team: The project ma

American project manager, there are twsubgroup is composed of three team memembers)

- J-A: Mixed cultural team: The project maproject manager, there are twp subgrou

- J: Japanese team: All team members haveA and J represent the single cultural teams,cultural teams. The Y axis shows hidden w

There are two types of quality risks as

risks and project quality risks.

There is no significant difference in fu

and mixed cultural teams. In an IC-VDT/VD

likelihood that components produced by the p

exception handling (Jin and Levitt 1996). No

behavior patterns do not affect any quality ris

Figure 7.3: Hidden Work Volume (A Organization Style): Case of ReciprocalWorkflow

ect outcomes, hidden work volume (person-s the combinations of team members:

e American cultural background nager is American. Underneath the o subgroups: American and Japanese. Each mbers (one sub leader and two sub team

nager is Japanese. Underneath the Japanese ps: American and Japanese. Japanese cultural background.

while A-J and J-A represent the mixed-ork volume (person-month).

outcomes of IC-VDT: functional quality

nctional quality risk between single cultural

T sense, the project quality risk refers to the

roject have defects based on rework and

difference implies that participants’

ks in technical defaults and defects.

142

The project quality risk for the Japanese organization style is about 0.3431 for

single cultural cases, A or J. The mixed cultural team cases with Japanese organization

style increased to 0.5. In the cases of the American organization style, the project quality

risk keeps showing the high level, around 0.5. Based on the first intellective experiment

(Chapter 5), the magnitude of differences due to organization styles (27-49%) is bigger

than that of differences in behavior patterns (3-21%). However, the increased project

quality risk (41-54%) nearly fills the gaps caused by changes in organizations styles.

This indicates the degree of impacts of having mixed cultural teams.

Functional Quality Risk

0.0

0.2

0.4

0.6

A J A-J J-A

Combinations of Team Members

A Org. Style

J Org. StyleProject Quality Risk

0.0

0.2

0.4

0.6

A J A-J J-A

Combinations of Team Members

A Org. Style

J Org. Style

Figure 7.5: Project Quality Risk: Case of

Reciprocal Workflow

Figure 7.4: Functional Quality Risk:Case of Reciprocal Workflow

Note: The figures illustrate functional quality risks and project quality risks for reciprocal cases. The X axis shows the combinations of team members: - A: American team: All team members have American cultural background - A-J: Mixed cultural team: The project manager is American. Underneath the

American project manager, there are two subgroups: American and Japanese. Each subgroup is composed of three team members (one sub leader and two sub team members)

- J-A: Mixed cultural team: The project manager is Japanese. Underneath the Japanese project manager, there are twp subgroups: American and Japanese.

- J: Japanese team: All team members have Japanese cultural background. A and J represent the single cultural teams, while A-J and J-A represent the mixed-cultural teams. The Y axis shows quality risks. Risks are scaled from 0 to 1.0 in VDT/IC-VDT.

31 VDT and CC-VDT scale both functional and project quality risks from 0 to 1.0. Based on SimVision ® 4.0.0 Help Files, any quality risk below 0.2 is probably acceptable, while risk greater than 0.5 indicates high quality risk.

143

2. Impacts of Matrix Strength: The intellective experiment examines the

effects of changing the level of matrix strength from low to high to see the magnitude of

influence on team outcomes.

Table 7.9 illustrates the outcomes measured as the averages of increased project

duration, work volume, or quality for all cases. High matrix strength can improve team

performance along the dimensions of project duration and hidden work volume. The

range of improvement is average around 14-16%. However, the changes in matrix

strength do not affect both product and project quality risks.

Table 7.9: Average of Increased Duration, Work Volume, or Quality

Average of increased duration, work volume, or quality

Project Duration 14.1% improved

Hidden Work Volume 16.6 % improved

Project/Functional

Quality Risk

No improvement

Note: This table illustrates the average impacts of changes from medium matrix strengthto high matrix strength under the same project complexity. The outcomes are measured as the averages of increased project duration, work volume, or quality for all cases.

*“Improvement” means shorter project duration, less hidden work volume, and less quality risk.

*Average of all cases

X 100Outcomes of high matrix strength Outcomes of medium matrix strengths *% =

T

cases of

changing

performa

high, pro

case of J

gradually

different

o be more precise, I compare the cases of Japanese organization style with the

American organization style. In the case of the American organization style,

the matrix strength from low to medium does not significantly improve

nce (Figure 7.6). However, when changing the matrix strength from medium to

ject duration is dramatically shortened and improved. On the other hand, in the

apanese organization style, project duration for all team combinations is

shortened (Figure 7.7). The other observation is that there is no distinguished

tendency between single cultural cases and mixed cultural cases.

144

Project Duration(A organization style)

30.0

40.0

50.0

60.0

70.0

L M H

Matrix Strength

A

J

A-J

J-A

Project Duration(J organization style)

30.0

40.0

50.0

60.0

70.0

L M H

Matrix Strength

AJA-JJ-A

Figure 7.7: Project Duration (J Organization Style): Case of Reciprocal Workflow

Figure 7.6: Project Duration (A Organization Style): Case of Reciprocal Workflow

Note: These figures demonstrate the impacts of changes in matrix strength. The X axis shows the matrix strength from low to high. The Y axis represents the project duration. The longer the project duration becomes, the worse the project performance will be. Therefore, improvement indicates shorter project durations when changing the matrix level from low to medium or from medium to high. The left graph is for reciprocal cases with the American organization style, while the right side is for reciprocal cases with the Japanese organization style.

Figures 7.8, 7.9, 7.10, and 7.11 show functional quality risks and project quality

risks. There is no significant difference in functional quality risk in any case. This

means that the changes in matrix strength do not improve any product quality risks in

technical defaults and defects.

In the case of the American organization style, project quality risks are improved

when the matrix strength changes from medium to high. In the case of the Japanese

organization style, project quality risks are not improved except for the J-A mixed

cultural case. The J-A mixed cultural teams with both the J and A organization styles

show greater improvement (6%-7%), when the matrix strength varies from medium to

high. Mixed cultural teams (A-J and J-A combinations) with the Japanese organization

style show high project quality risks as compared with single cultural teams (A and J

teams).

145

Functional Quality Risk(A organization style)

0.3

0.4

0.5

0.6

L M H

Matrix Strength

A

J

A-J

J-A

Functional Quality Risk(J organization style)

0.3

0.4

0.5

0.6

L M H

Matrix Strength

A

J

A-J

J-A

Figure 7.9: Functional Quality Risk (J Organization Style): Case of Reciprocal Workflow

Figure 7.8: Functional Quality Risk (A Organization Style): Case of Reciprocal Workflow

Project Quality Risk(A organization style)

0.3

0.4

0.4

0.5

0.5

0.6

L M H

Matrix Strength

A

J

A-J

J-A

Project Quality Risk(J organization style)

0.3

0.4

0.5

0.6

L M H

Matrix Strength

A

J

A-J

J-A

Figure 7.11: Project Quality Risk (J Organization Style): Case of Reciprocal Workflow

Figure 7.10: Project Quality Risk (A Organization Style): Case of Reciprocal Workflow

Note: These figures demonstrate product and project quality risks as outcomes of changes in matrix strength. The X axis shows the matrix strength from low to high. The Y axis represents the degree of quality risks. VDT/IC-VDT scale risks from 0 to 1.0. The greater the number becomes, the greater the risk will be. Therefore, improvement indicates the decreasing risk when changing the matrix level from low tmedium or from medium to high. The left graph is for reciprocal cases with theAmerican organization style, while the right side is for reciprocal cases with theJapanese organization style. Reciprocal cases are used as ex

o

amples.

146

The Impacts of Organization Style: The third objective of this experimentation

is to see the impacts of changes in organization styles. Specifically, there are two

different set of organization styles: divergence vs. convergence styles. The American

organization style represents divergence, while the Japanese organization style represents

convergence.

The changes in organization style affect project duration, hidden work volume,

and project quality risks (Figure 7.2-7.11). Based on Figures, the Japanese organization

style shows that mixed cultural teams have consistently greater coordination costs —

increased hidden work volume, increased project duration and project quality risk — than

the single cultural teams. On the other hand, the American organization style results have

no differences in coordination costs, even though some cases show better performance32.

Figure 12 demonstrates the variance of increased duration when changing team

combinations from a single culture (J or A teams) to a mixed culture (J-A or A-J

combinations). Increased project duration indicates that a mixed cultural team shows less

efficient performance in duration, while decreased duration indicates that a mixed

cultural team is more efficient. Then, I compare the variance of American organization

cases with that of Japanese organization cases. The range of American organization

cases is between -20% and +40% with 6.3% as means. In the case of Japanese

organization cases, the range is bigger and wider than that of American organization

cases, between -30% and +76% with 7.7% as means. Several implications are considered

as follows:

First of all, this result indicates that mixed cultural teams can be less efficient (e.g.,

>70% worse) or slightly more efficient (e.g., 30% improved) than single cultural cases.

The second is that the A organization style (decentralized structure) has less

negative-impacts of having mixed cultural teams than the J organization style

(centralized structure). The J organization style (centralized structure) can be more

efficient (improved by 30%) or be less efficient (increased by 76%) than the A

organization style. This might be the main reason why American practices are used and

diffused in the international business scene. Additionally, one speaker from World Bank

32 Figure 7.12 shows that there are several cases where mixed-cultural teams have better performance than single cultural teams. This tendency is observed mostly in the sequential workflow cases and/or in the cases of the American organization style.

147

at the CRGP summer program mentioned “decentralization is definitely the key to

leading a global team.” Therefore, emergent simulated results potentially support that

American practices, divergence styles, are used and diffused in the international scene

because of less negative impacts of having mixed cultural teams based on the view

through information processing lens.

Finally, these results may answer the question of why mixed-cultural teams can be

either more or less effective than single-cultural teams. Some researchers have pointed

out that mixed cultural teams become productive when well managed (Fiedler, 1966;

Kumar et al, 1991). On the other hands, other researchers describe difficulties and

conflicts in managing global projects (Beamish, 1985; Cullen et al, 1995). The range

between less efficient and more efficient duration that they observed might potentially

represent the outcomes of random sampling from real IJV projects.

Variance of duration: Means +6.3%, StDev 14% American Style

Decentralized authority -20% +40% (intensive case) Medium formalization Flat hierarchy

Means +7.7%, StDev 18% +76% (intensive case)-30%

Japanese Style Centralized authority

High formalization Multi-level hierarchy

80%60% 40% 50% 70% -20% 20% 30%0%-30% 10%-10%More efficient Less efficient

Mixed cultural teams

Outcomes of mixed cultural teams *% = Outcomes of single cultural teams X 100

Figure 7.12: Changes in Duration for Mixed-Cultural Teams

Note: This figure demonstrates the changes in project duration when changing team combinations from a single cultural case (J or A) to a mixed cultural case (A-J or J-A). The X axis shows increased or decreased duration. Plus, “+,” indicates that mixed cultural teams are less efficient. Minus, “-,“ means that mixed cultural teams are more efficient than single cultural teams.

The first intellective experiment (Chapter 5) concludes that the impact of changes

in organization styles is greater than the changes in micro-level behavior patterns.

However, using mixed cultural team members can have a larger impact than having the

148

wrong organization style. For example, the project quality risk of the J-A case jumped up

to 0.5 from 0.34 in the cases of single cultural teams (Figure 7.5). 0.5 is much closer to

the levels achieved in the American organization style. Therefore, changes in team

members can have equal impact to changes in organization styles.

3. Impacts of Institutional Exceptions: One of the new features in IC-

VDT is a model of the concept of institutional exceptions. Based on current research

(Orr, 2004; 2005; Mahalingam, 2004; 2005), institutional exceptions arise when two

cultures have large differences in institutional and cultural systems. In these experiments,

the probability of institutional exceptions is set to 1.0 for all cases. This means that

institutional exceptions will happen, given a project exception, with the same probability

as the project exception occurring. So the probability of an institutional exception is the

square of the probability of a project exception. This dissertation focuses on qualitative

differences and performance dynamics, rather than quantitative accuracy. However, it is

worth checking how many institutional exceptions are generated and how much impact

they have on performance. IC-VDT models three types of exceptions: functional, project,

and institutional. Table 7.10 illustrates the ratio of institutional exceptions to total

exceptions. The greatest percentage of institutional exceptions is only 4.3% of total

exceptions. However, the impact of these exceptions for mixed-cultural teams is far

greater than four percent. This implies that small difference in the number of institutional

exceptions causes large outcome differences in some situations. A more accurate

quantitative analysis is required as the next step.

Table 7.11: Ratio of Institutional Exceptions to Total Exceptions

Low Medium High

Pooled 0% 0% 0%

Sequential 1.6% 1.8% 0.9%

Reciprocal 4.3% 3.7% 1.5%

Intensive 3.2% 2.5% 1.5%

149

7.3.2 Statistical Analyses

Howell suggests the use of the Pearson correlation and the probability value test for

continuous variables (Howell, 2002) to measure the strength of a relationship between

two variables. The probability value, called p-value, assesses the significance of a

relationship between two variables. The strength of the relationship is indicated by the

correlation coefficient, r. The significance of the relationship is expressed in probability

levels: p (e.g., significant at p =.05), indicating the likelihood that a given correlation

coefficient, r, occurs by chance, given no relationship in the population. Therefore, the

smaller the p-level, the more significant the relationship.

In this experiment, I address four main hypotheses:

- Hypothesis 1: The greater the project complexity, the greater the quantity of

information the project team has to handle

- Hypothesis 2: A divergence approach (American organization style) shows

better performance for mixed cultural teams than a convergence approach

(Japanese organization style)

- Hypothesis 3: A mixed cultural team needs to handle a greater quantity of

information than a single cultural team

- Hypothesis 4: High matrix strength can improve the team performance of

mixed cultural teams

I use the following independent and dependent variables for the hypothesis

testing:

- Independent variables: degrees of task complexity (pooled, sequential,

reciprocal, intensive), types of organization styles (divergence style,

convergence style), types of team members (single cultural, mixed cultural),

and matrix strength (low, medium, and high)

- Dependent variables: project duration, cumulative hidden work volume,

functional quality risk, and project quality risk.

I test the strength and significance of the relationships between dependent and

independent variables. In the following tables, the significance levels are marked as

follows: p < 0.10 (†), p < 0.05 (*), p ＜ 0.01 (**). Two asterisks, **, indicates a highly

significant finding. The total number of experiments is 96 cases (Table 7.1) and is larger

150

than the minimum number for doing hypothesis testing recommended by Howell (2002).

I begin by looking at the entire set of simulated results (Table 7.11), and then examine the

strength and significance of relationships between the variables at each task complexity

level (Tables 7.12 – 7.15).

Table 7.12: The Correlation Coefficient (r) for all cases

All Cases Project Duration

Cumulative Hidden Work

Volume

Functional Quality

Risk Project

Quality Risk

1

Type of interdependency(Pooled, Sequential, Reciprocal, Intensive flow)

0.48** 0.66** -0.08 0.75**

2

Type of organization (A-vs.-J: Centralization, Hierarchy, Formalization)

-0.19† -0.29* 0.01 -0.16

3

Type of members (Single cultural vs. Mixed cultural teams)

0.04 -0.03 0.01 0.18†

4 Matrix Strength (Low, Medium, High) -0.22* -0.29* 0.01 -0.02

s

v

c

t

Note: This table shows the correlation coefficient, r, between independent variables anddependent variables. The p-level is labeled: † p < 0.10, * p < 0.05, ** p < 0.01. Total sample size: n=96

Table 7.11 indicates that the level of project complexity has a significant and

trong relationship with project team outcomes such as project duration, hidden work

olume, and project quality risk, since the p-values are below 0.01 and the correlation

oefficients are above 0.48. However, there is no impact on functional quality risk. The

ype of organization (A or J) has a large impact on project duration and hidden work

volume, but not functional and project quality risks. The combinations of team members

(J, A, J-A, or A-J) are weakly related to project quality risk. The level of matrix strength

improves performance on project duration and hidden work volume, but not functional

and project quality risks. Overall, the level of task complexity has the biggest impact on

team performance, as predicted by contingency theory. This is reasonable since VDT has

151

been developed based on contingency theory. In other words, this intellective experiment

qualitatively confirmed internal validity.

The secondary variables are organization-style-related settings such as divergence

vs. convergence styles and matrix strength. These results confirm hypotheses proposed

by Adler (1997) and Hofstede (1991) to a limited degree. Both the organization-style-

related settings show strong correlation with project duration and hidden work volume,

but not with functional and project quality risks.

Finally, the types of team members (single cultural vs. mixed cultural) show

relatively weaker relationships with the outcomes. Roughly, the order of significance of

these variables is (1) project complexity, (2) organizational settings (type of organization

and matrix strength), and (3) type of members.

Next, I look at relationships for each level of task complexity such as pooled,

sequential, reciprocal, and intense workflows, because task complexity is generally given

and relatively fixed in a real-life situation (Table 7.12-15). I also test types of team

members with American (divergence) organization style and with Japanese

(convergence) organization style.

Table 7.13: The Correlation Coefficient (r) for Pooled Workflow

Pooled Workflow Project

Duration

Cumulative Hidden Work

Volume

Functional Quality

Risk

Project Quality

Risk

1

Type of organization (A, J: Centralization, Hierarchy, Formalization)

-0.63** -0.77** -0.02 -

2

Type of members (Single cultural vs. Mixed cultural teams)

0.66** -0.21 0.05 -

A org. style (n=12) 0.85** -0.41 -0.11 -

J org. style (n=12) 0.86** -0.25 0.20 -

3 Matrix Strength (Low, Medium, High) 0.32 0.44* -0.04 -

Note: This table shows the correlation coefficient, r, between independent variables and dependent variables. The p-level is labeled: † p < 0.10, * p < 0.05, ** p < 0.01. Total sample size: n=96

152

Table 7.14: The Correlation Coefficient (r) for Sequential Workflow

Sequential Workflow Project

Duration

Cumulative Hidden Work

Volume

Functional Quality

Risk

Project Quality

Risk

1

Type of organization (A, J: Centralization, Hierarchy, Formalization) 0.12 -0.18 -0.05 -0.70**

2

Type of members (Single cultural vs. Mixed cultural teams) 0.18 0.01 0.10 0.57**

A org. style (n=12) 0.19 0.07 0.03 0.66*

J org. style (n=12) 0.32 -0.39 0.17 0.92**

3 Matrix Strength (Low, Medium, High) -0.62** -0.53** 0.04 -0.05

Note: This table shows the correlation coefficient, r, between independent variables and dependent variables. The p-level is labeled: † p < 0.10, * p < 0.05, ** p < 0.01. Total sample size: n=96

Table 7.15: The Correlation Coefficient (r) for Reciprocal Workflow

Reciprocal Workflow Project

Duration

Cumulative Hidden Work

Volume

Functional Quality

Risk

Project Quality

Risk

1

Type of organization (A, J: Centralization, Hierarchy, Formalization) -0.59** -0.76** 0.12 -0.49*

2

Type of members (Single cultural vs. Mixed cultural teams) 0.36† 0.10 -0.03 0.70**

A org. style (n=12) 0.40 -0.12 0.06 0.58*

J org. style (n=12) 0.53* 0.48† -0.12 0.98**

3 Matrix Strength (Low, Medium, High) -0.50** -0.47* -0.01 -0.15

Note: This table shows the correlation coefficient, r, between independent variables and dependent variables. The p-level is labeled: † p < 0.10, * p < 0.05, ** p < 0.01. Total sample size: n=96

153

Table 7.16: The Correlation Coefficient (r) for Intensive Workflow

Reciprocal Workflow Project

Duration

Cumulative Hidden Work

Volume

Functional Quality

Risk

Project Quality

Risk

1

Type of organization (A, J: Centralization, Hierarchy, Formalization) -0.58** -0.51** 0.02 -0.53**

2

Type of members (Single cultural vs. Mixed cultural teams) 0.02 -0.14 -0.07 0.72**

A org. style (n=12) -0.00 -0.24 -0.07 0.67*

J org. style (n=12) 0.07 -0.06 -0.08 0.99**

3 Matrix Strength (Low, Medium, High) -0.64** -0.63** 0.05 0.03

Note: This table shows the correlation coefficient, r, between independent variables anddependent variables. The p-level is labeled: † p < 0.10, * p < 0.05, ** p < 0.01. Total sample size: n=96

When the task complexity is fixed, organization settings such as matrix strength

and organization styles clearly affect project outcomes significantly. In particular, the

type of organization style influences the project duration, cumulative hidden work

volume, and project quality risk, confirming Adler’s hypothesis (Adler, 1997). The

convergence style is less efficient for mixed cultural teams than the divergence style.

Matrix strength is significantly related to project duration and hidden work volume. In

the case of intensive workflow, this tendency is amplified.

The type of team members shows a relationship with project quality risk overall.

However, there is no consistency in project duration and hidden work volume. In some

situations (pooled workflow), the type of team members significantly influences project

duration, but not in other situations (sequential and intensive workflows).

When comparing between American and Japanese organization cases, Japanese

organization cases show relatively clear relationships — i.e., the correlation coefficients

in Japanese organization are double those for the American organization style. This

indicates that mixed cultural teams are relatively less efficient in the case of Japanese

organization style than in the case of American organization style. In other words, under

154

the American organization style, mixed cultural teams do not have to deal with greater

coordination costs, as compared to single cultural teams. One possible reason is that the

low level of centralization and the flat hierarchy (A organization style) mitigates the

impact of institutional costs, as Adler has proposed (1997).

7.4 Discussion and Conclusion In this chapter, I have described an intellective experiment that examines the impacts of

mixed cultural teams on project performance, and that looks at, more ore less, the

effective management and leadership styles in IJV projects. The simulated results are

statistically compared to the four hypotheses I proposed in this chapter. This section

discusses the four following topics: validation, limitations, implications and conclusion.

7.4.1 Implications

I discuss three topics: (1) mixed-cultural teams vs. single cultural teams, (2) effective

organization styles in global projects, and (3) reasons for conflicts and struggles in IJV

projects.

(1) Mixed- vs. Single-cultural teams: Research findings from Beamish (1985)

and Cullen et al (1995) exemplify that mixed cultural cases have greater coordination

costs than single cultural cases. This indicates that mixed cultural teams need to handle

large amounts of information during a project, based on the information processing view.

A statistical analysis of the simulated results confirms this presumption under limited

conditions. For instance, in the case of the Japanese organization style, mixed cultural

teams show increased hidden work volume and long project durations. However, in the

case of the American organization style, this tendency vanishes — some mixed cultural

cases show actually better performance than single cultural cases. Adler (1997) proposed

that mixed cultural teams could potentially become the most effective and the least

productive, as shown in Figure 7.13. When comparing Figure 7.13 with the simulated

results (Figure 7.12), both figures showed that the range of outcomes for mixed cultural

teams is wider than for single cultural teams. For instance, the mixed cultural teams can

be more effective (30% improvement), or can be less effective (76% less efficient).

155

Therefore, the simulated results from this intellective experiment support Adler’s

hypothesis.

Single Cultural Teams Highly ineffective

Mixed-Cultural Teams Highly effective Figure 7.13: Team Effectiveness

Note: The figure demonstrate the conceptual distribution of team effectiveness in mixed- vs. single-cultural teams. This figure is adapted from Adler (1997, pp.137) and Kovach (1976). The X axis represents team performance. They argued that mixed cultural teams tend to have a wider range of outcomes than single cultural teams.

Levitt argues that increased coordination costs are caused by both project

complexity and institutional/cultural complexity in global projects (Levitt, 2004). These

intellective experiments focus on and represent only two cultural cases, Japanese and

American cultures. Bi-cultural cases (A-J or J-A) still show higher coordination costs

than single cultural cases (A or J) — i.e., a mixed-cultural team shows 70% longer in

project, compared to a single cultural team. In most global projects, more than two

cultures are generally involved, implying greater institutional costs than my simulated

results with just two cultures. The combination of institutional costs and coordination

costs can overwhelm IJV teams, contributing to the high failure rates of IJV projects —

i.e., Beamish (1985) found that 50% of IJV projects have some struggles and/or conflicts.

156

$, Time, Efforts etc

Institutional costs

Coordination costsInstitutional Complexity

(Multiple, different cultures and institutions) Direct costs

Coordination Complexity (Task, uncertainty, and interdependencies)

Figure 7.14: Institutional Costs and Coordination Costs

Note: The figure that is adapted from Levitt (2004) demonstrates the increased coordination costs affected by both coordination complexity and institutional complexity. Levitt argued that the greater the project complexity becomes, the greater the coordination costs project teams need to consider (X axis). Similarly, higher institutional complexity leads to greater additional coordination costs, called institutional costs (Y axis). If a project involves high coordination complexity with high institutional complexity, total coordination costs including institutional costs become very high, for exceeding direct costs.

It is important to discuss why mixed cultural teams usually perform less

effectively than single cultural teams and, also, why they are sometimes more effective.

IC-VDT incorporates institutional exception mechanisms (Chapter 6). Based on my case

studies and my CRGP colleagues’ findings, I started designing and calibrating the model,

so that institutional exceptions increase work volume by about 5%. Thus, it does make

sense that the average of increased duration and hidden work volume is around 2-7%.

However, the wide ranges of impacts for mixed cultural teams (between 76% increased

project duration and 30% improved project duration) were unexpected (Figure 7.12).

Why did mixed-cultural team perform so much worse than expected? Why was project

157

duration increased by 76%? One possible reason is related to the turbulent in

information flow point discussed by Fyall (2001). The turbulent point refers to a point

at which the change in slope is greatest, circled on Figure 7.15. The analogy here is to

the Reynolds Number in fluid mechanics, where laminas flow changes to turbulent flow,

with head loss proportional to the square of the velocity (vs. the velocity)

600 Transition Laminar Turbulent

Institutional exceptions make the turbulent point at lower error probabilities

Indi

rect

Wor

k (D

ays)

100 Turbulent point

30 0.3 0.1 Error probability

Figure 7.15: Turbulent Point and Institutional Exceptions

Note: The figure that is adapted from Fyall (2001) demonstrates the “turbulent” point for a VDT organization. The logarithmic graph shows the changes in slope corresponding to error probability (X axis). When error probability reaches a certain point, indirect work will increase dramatically and exponentially. Institutional exceptions make mixed cultural teams more sensitive to error probability, thus project uncertainty.

In IC-VDT, institutional exceptions substantially increased total error probability,

making a turbulent point earlier than usual. Therefore, mixed cultural teams may become

more sensitive to project uncertainty. In other words, even if single cultural teams are not

affected by increasing project uncertainty slightly, mixed-cultural teams may need to

cope with large indirect work volume.

The simulated results also showed some cases in which mixed cultural teams

performed better than single cultural teams. As one possible interpretation, in the case of

sequential workflow, institutional exceptions may not be critical factors, because actors

158

have low backlogs. No parallel activity gives actors enough slack time to handle the

institutional exceptions.

Secondly, a combination of J type and A type can affect the demand for

information processing. Japanese actors tend to require or to give a larger volume of

information and have a high probability of attending to communications. These

tendencies increase total demand for information processing, sometimes overwhelming

the capacity of a project team. However, when A type actors joined the J type actors, the

demand for information processing is reduced, resulting in positive outcomes.

The third possibility is a random combination of actors, task interdependencies

and organization styles. The three factors are interacted each other, creating complex

organizational behaviors. For some combinations of the factors, the performance of

mixed cultural teams becomes better. Sensitivity analysis is suggested to diagnose where

the tipping point to positive outcomes is located.

(2) Effective Organization Style: Finding an effective leadership style for

global projects has been an unanswered question, at least theoretically, leaving a space to

investigate and theorize about it. Adler (1997) argued that the divergence style is

preferred for mixed cultural and culturally diverse teams. The statistical analysis and

simulated results confirm this hypothesis. There are several possible explanations for the

model results. At first, low centralization reduces the frequency at which a worker passes

exceptions for resolution to his/her supervisor, substantially reducing institutional

exceptions. With high centralization, the opposite is true. The second reason is that a flat

hierarchy has a similar effect to a low level of centralization, since there are fewer levels

above any given actor that can generate institutional exceptions due to cultural

differences.

159

Figure 7.16: American Organization Structure Type with Interfaces

Note: The above figures illustrate exception handling structures of American and Japanese styles. Actors colored green represent J culture and the other actors colored yellow represent A culture. The highlighted lines indicate interfaces in which different cultural actors are involved. In the case of Japanese organization style, centralization requires that subordinates report their exceptions directly to the project manager, expressed by the doted line. Therefore, the frequency of institutional exceptions is increased in the Japanese organization structure, while the American organization structure has few, because of a decentralized structure with a flat hierarchy.

Figure 7.17: Japanese Organization Structure Type with Interfaces

The matrix swing concept from Morris (1982) argues that a project’s task

complexity (in terms of interdependence between tasks) changes systematically over the

life cycle of a project. He argues that the organization needs to adapt accordingly.

Construction projects generally begin with a highly centralized and small team, with a

high degree of reciprocal interdependence. At this stage, the emphasis is on judgment

about technical performance and business feasibility. During the implementation stage,

construction projects usually formalize a large team with a decentralized structure,

because various task requirements are now clear. Sequential interdependence is typical at

this stage. The final stage is the turnover phase that requires final testing and punch lists.

During the turnover phase, a project team becomes medium size with a centralized

structure in order to integrate all components and hand off the project to an owner. This

phase has sequential and/or reciprocal interdependences.

The matrix swing concept from Morris (1982) argues that a project’s task

complexity (in terms of interdependence between tasks) changes systematically over the

life cycle of a project. He argues that the organization needs to adapt accordingly.

Construction projects generally begin with a highly centralized and small team, with a

high degree of reciprocal interdependence. At this stage, the emphasis is on judgment

about technical performance and business feasibility. During the implementation stage,

construction projects usually formalize a large team with a decentralized structure,

because various task requirements are now clear. Sequential interdependence is typical at

this stage. The final stage is the turnover phase that requires final testing and punch lists.

During the turnover phase, a project team becomes medium size with a centralized

structure in order to integrate all components and hand off the project to an owner. This

phase has sequential and/or reciprocal interdependences.

This life cycle pattern of construction projects helps us understand which stage is

more critical, and how to change organization styles in accordance with a project in

progress. The American organization style that has a decentralized structure implies

fitting into the implementation stage. On the other hand, the Japanese organization style

with a centralized structure is most likely better at the beginning or at the end. When

This life cycle pattern of construction projects helps us understand which stage is

more critical, and how to change organization styles in accordance with a project in

progress. The American organization style that has a decentralized structure implies

fitting into the implementation stage. On the other hand, the Japanese organization style

with a centralized structure is most likely better at the beginning or at the end. When

160

overlapping this to mixed-cultural team cases, mixed-cultural teams have large potential

to show struggles at the beginning or at the end, because construction projects demand a

centralized structure in order to integrate opinions and judgment, and because high

reciprocal interdependence cause a high demand for information processing. Anecdotes

acquired through the CRGP General Council Round Table conference (2005) and

informal interviews during my research support the assumption that mixed-cultural teams

have failed to integrate their opinions or to make a consensus among participants during

the project shaping phase (Millar and Lessard, 2000) or the turnover phase (Morris, 1982).

Further investigation is required to prove this assumption.

(3) Matrix strength: High matrix strength can improve the performance of

mixed cultural teams (Hofstede, 1991). Davis and Lawrence stressed geographical

connectedness as a key success factor in managing global and transnational corporations

(Davis and Lawrence, 1973). The simulated results indicate that high matrix strength

improves team performance in terms of project duration and hidden work volume.

However, matrix strength does not influence functional quality risk or project quality

risks. In IC-VDT, increased emphasis on informal communication yields better outcomes

for a co-located (high matrix strength) project team.

(4) Bridging the Gap between the Real World and the Simulation: Why do

so many international joint venture projects show poor performance? The simulated

results suggest several possible explanations. The first possibility is that joint venture

projects do not set up an appropriate organization style to address their levels of project

complexity and task interdependency. For instance, if a PM insisted on using his/her

traditional style to manage and control a project organization, his/her style may not fit

into a given project situation, causing misfits. The second possibility is that a project

team does not set up an appropriate organization style that fits team members who come

from different countries. Participants from different countries tend to have differing

behavior patterns. A combination of behavior patterns may cause unexpected

inefficiency. Even though the differences in behavior patterns are small, the aggregated

cost can be large. Finally, project participants may be unaware of institutional exceptions

161

caused by differing task control styles. Lack of individual or group knowledge and

experience in international projects may amplify and exaggerate the impacts of

institutional exceptions. Therefore, there are three possible components of increased

coordination costs: (1) misfits between project complexity and organization style, (2)

misfits between individual behavior patterns and organization style, and (3) lack of

individual/group experience in cross-cultural teams.

7.4.2 Validation

Validation has been a challenging problem in computational organization theory research.

I adopt an established framework (Thomsen et al, 1999) for validating simulation models

of organizations. The evaluation trajectory proposed by Thomsen et al specifies a

strategy for building up successive validation experiments for new models (Figure 3.1).

Thomsen discussed three major steps of validation: reasoning, representation, and

usefulness. As the first step, the reasoning assumptions of the simulation model must be

validated (Thomsen et al, 1999) and is the main goal of validating IC-VDT. Specifically,

micro-theories relating to observable micro-behavior must match the behaviors observed

in the simulation. In the IC-VDT model, I create a set of micro-level behavior patterns

by linking cultural theory to observed micro-level behavior through ethnographic

interviews (Chapter 4). To validate the interactions between organization theory axioms

and the emergent macro-behavior of my simulation model, I use intellective simulation

experiments of an idealized organization (Chapter 5). This intellective simulation step

enables me to validate the J and A micro-level behavior patterns. The qualitative

consistency between the simulated results and existing literature suggests that the micro-

level behavior patterns of the two cultures are appropriately encoded in the model. The

second step is to validate the reasoning assumptions of the simulation model for mixed-

cultural teams (Chapter 7). Specifically, IC-VDT addresses institutional exception

mechanisms. I conducted a second intellective experiment to validate my assumptions

about institutional exceptions.

162

Macro-Theory Simulation Observation

Theoretical consistency?

Computational Model Output

Cultural elements Idealized Project Outputs 1) Practices: Organization styles Run Simulation

Macro-TheoryTask control styles VDT / IC-VDT 2) Values:

Define Inputs Idealized Project Inputs Micro-level behaviors

Figure 7.18: Validation of the Reasoning Assumptions of IC-VDT

Note: This figure shows the detailed steps for validation of the reasoning assumptions of IC-VDT. Observations through case studies reveal distinctive cultural elements along value and practice dimensions. Organizational macro-theory can specify both idealized project inputs — idealized configurations of work processes and task complexities — and idealized project outputs — idealized project outcomes of duration and work volume. Both observed cultural elements and idealized project inputs define the inputs of simulations. Finally, model predictions are compared to idealized project outputs, confirming theoretical consistency.

Figure 7.18 illustrates the detailed steps for validation of the reasoning

assumptions of IC-VDT. As the first step, I encoded cultural elements observed during

case studies along value-practice dimensions. Both observed elements and idealized

project inputs defined the inputs of simulations. The second step is to compare model

predictions to idealized project outcomes that are based on empirical findings in the

micro- and macro- organization theory and culture literature. The purpose of this

approach is to confirm theoretical consistency. This dissertation completed two cycles of

this validation: the first intellective experiment for single cultural cases and the second

intellective experiment for mixed cultural cases. The two intellective experiments

successfully validated the reasoning of IC-VDT acquired organizational and cultural

contingency theory, and added some refineries — convergence vs. divergence leadership

style and matrix strength — as propositions to be tested in the research.

Future research should attempt to validate the representation and usefulness of IC-

VDT. The second component of the Thomsen’s validation trajectory, representation,

would assess the ability of IC-VDT to model and represent a real organization in terms

that are meaningful to managers. The third component of the trajectory, usefulness,

would validate the applicability of IC-VDT to real global organizational scenarios,

163

initially through “gedanken” or thought experiments in which IC-VDT outputs are

compared to managers’ predictions, and then through natural experiments performed

retrospectively, concurrently and finally, prospectively.

7.4.3 Limitations

A computational simulation has known limitations in applying it to a real situation.

Needless to say, IC-VDT models and captures key cultural and institutional phenomena

that are observed in IJV projects, but not include all social and institutional factors. In

particular, there are four main limitations of IC-VDT.

IC-VDT is limited to only two cultural cases (Japanese and American cultures).

Although focusing on two cultural cases is an excellent strategy to cast the distinctive

cultural and institutional factors in relief, it is hard to find an IJV project that is

assembled from just two specific cultures. However, IC-VDT can readily generate sets

of micro-level behavior patterns for other cultures based upon Hofstede’s research (1991).

This function allows future researchers to apply IC-VDT to other cultural cases.

This research assumed that participants’ motivation and productivity is consistent.

Psychological factors such as motivation, emotion, and trust can influence on

productivity. From the case study, an American engineer, for instance, lost his

motivation when he had to follow Japanese business customs. An interviewee mentioned

that some expatriates quit a project at the middle of a project because of low motivation

or because of lack of trust with their partners or colleagues. The common opinion from

interviewees is just reluctant to do unknown or unfamiliar practices even though they

know neither what real outcomes are, nor what kinds of advantages they can get.

Computational simulations such as IC-VDT have the potential to show what if they adapt

partner’s approach, resulting in mitigating over reaction against using unfamiliar

practices.

This research makes the assumption that team members do not adapt their values

or practices during the course of a given project. However, researchers have increasingly

been interested in how people learn cultural values and cultural practices from each

other (e.g., Orr, 2004; 2005). In addition, corporations can externalize their own tacit

knowledge into explicit knowledge based upon global experience, reducing and

164

minimizing negative impacts on project performance caused by cultural and institutional

complexity. Institutionalizing global knowledge and experience — e.g., setting up a rule,

guidance, or process for global projects — can provide another solution to mitigate

problems observed in IJV projects.

Unfortunately, there exists no strong theory to support our proposition about

institutional exceptions caused by differing institutionalized practices. Many institutional

researchers have focused on the differences or dynamics of institutions. Very few have

investigated conflicts caused by institutional differences, except for studies of long-term

outcomes (Aoki, 2001). This calls for researchers to carry out further qualitative and

quantitative analysis, and to theorize about the concept of institutional exception

mechanisms.

In addition, the current research does not take into consideration the potential

positive impacts of cultural interactions, such as innovation, creativity, and the sharing of

advanced technology. A few researchers have started exploring innovation issues in

project-based organizations (e.g., Taylor and Levitt, 2005; Cox and Blake, 1991).

Furthermore, trust relationships need to be further explored. As many researchers

have pointed out (e.g., Ouchi, 1981; Zolin, 2002), developing a trust relationship at the

early stages of global projects is a key factor in enhancing project outcomes.

Finally, exogenous factors such as political and economic environments were

controlled for in this study. There is a need to integrate both endogenous and exogenous

factors influencing team performance. Furthermore, this research focuses on culturally-

driven normative systems in IJV project teams. It needs to incorporate not only other

elements of normative systems — i.e., professional and organizational cultures, but also

regulative and cultural-cognitive systems.

7.4.4 Conclusions

In this chapter, I have described an intellective experiment that compared simulated

results with theoretical outputs in order not only to understand and analyze the effects of

mixed-cultural teams, but also to seek a better organization style for mixed cultural teams.

The results demonstrated consistency with the starting hypotheses of this research to a

limited degree. Four hypotheses were examined.

165

The first hypothesis was linked to contingency theory (Thompson, 1967). Both

VDT and IC-VDT were developed and calibrated based on contingency theory. To

confirm this theory through intellective experimentation means not only that the project

description in the intellective experiment is robust and appropriate, but also that IC-VDT

conforms to contingency theory for the internal validity.

The second hypothesis was related to organization styles, decentralized structure

vs. centralized structure. The statistical analysis and simulated results confirmed the

hypothesis argued by Adler (Adler, 1997) that the decentralized style is preferred for

mixed cultural and culturally diverse teams.

The third hypothesis was linked to team combinations, — i.e., single cultural

teams vs. mixed cultural teams. Adler (1997) and Levitt (2004) argued that mixed

cultural teams can potentially perform less efficiently or more efficiently than single

cultural teams. Simulated results show an interesting variance of increased duration,

supporting this hypothesis. My statistical analysis indicated that the significance of this

hypothesis in terms of project duration and hidden work volume outcomes was not strong,

compared to other hypotheses. However, this hypothesis was strongly correlated to

project quality risks.

Finally, the intellective experiment confirmed the hypothesis that was linked to

matrix strength as a key management style for global projects and corporations. Matrix

strength significantly affected project duration and hidden work volume, but not project

quality risks.

To summarize, this intellective experiment confirmed that cultural and

institutional differences, culturally-driven normative systems, do influence team

performance in various ways. The American organization style with the high matrix

strength is apparently a better organizational practice for mixed cultural teams in general.

However, the Japanese organization style also has pros and cons. Project managers need

to design their project organization style by considering three factors — project

complexity, organization design, and participants’ behaviors — plus which stage a

project is currently in.

166

These conclusions are limited by a series of assumptions such as consistent

psychological factors, the specific cultural cases, no changes in behaviors over a project,

a linear learning curve, and no innovative processes.

167

CHAPTER EIGHT: CONCLUSION AND

CONTRIBUTIONS

The trend of globalization in the construction industry creates unique management

challenges. My dissertation focuses on understanding and analyzing the effects of

culturally-driven normative systems on project performance. The preceding chapters

have presented my ethnographies, the intellective experiment for single-cultural cases,

the extensions of VDT (IC-VDT), and the intellective experiment for mixed-cultural

cases. My ethnographies find distinct cultural factors between Japanese and American

firms along cultural value-practice dimensions. The intellective experiment for single

cultural cases shows the potential to incorporate cultural phenomena in a simulation

model with the information processing approach. My IC-VDT model allows managers

and researchers to seek a better organization design for mixed cultural teams. The

intellective experiment for mixed cultural teams supports empirical findings and

assumptions for managing global projects. The contents of the preceding chapters have

been published as several autonomous journal articles. This chapter concludes this

dissertation, and includes contributions to related disciplines and a discussion of future

research.

8.1 Conclusion Together with rapid advances in technology, the trend of globalization in the construction

has made facility engineering projects increasingly challenging, both with respect to the

efficiency of project executions and with respect to the effectiveness of project outcomes.

Research on international joint-venture (IJV) teams reveals high failure rates for projects,

due to difficulties in managing mixed-cultural teams. This dissertation attempts to

understand, analyze, and assess how cultural differences in IJV teams affect project

performance, using case studies and computational experimentations. In particular, this

dissertation focuses on two cultures — Japanese and American — as an example of the

minimum dyadic unit of cultural interaction in global construction projects. This

168

dissertation views “culturally-driven normative systems” as a key element of cultural

differences. Culturally-driven normative systems are composed of the two elements:

cultural values and cultural practices. Cultural values refer to the preferred behavior

patterns that people show when making task-related and communication-related decisions

in business situations. Cultural practices include cultural norms for adopting or using

specific organization designs to manage organizations and tasks.

This research used a research methodology based on the complementary

approaches of direct observations and computational experiments (Klhar and Simon,

2001). As the first step, I conducted direct observations with ethnographic interviews that

have potential for detecting new phenomena and for observing subtle information from

interviewees. Observed data were qualitatively characterized and analyzed through a

grounded-theory approach (Eisenhardt, 1989), such as analytic induction (Glaser and

Strauss, 1967) and cross-case pattern search (Yin, 1984). The second step was to use the

VDT computer simulation model as a “virtual laboratory” where we can address a series

of “what-if” questions (Dooley, 2002; Burton, 2003; Carley, 1995; 1996) in order to

understand and analyze the pure effects of cultural differences on team performance. This

research conducted two sets of intellective experiments: one for single cultural team cases

and one for mixed cultural team cases. The intellective experiments addressed and tested

relevant macro-organization theories: organizational contingency theory, cultural

contingency theory. And it explored the effectiveness of alternative leadership styles —

i.e., divergence vs. convergence styles and high vs. low matrix strength — for mixed

cultural teams, based on the literature survey. I compared the emergent macro-

organization simulation outcomes against predictions of macro-organization theory in the

sets of intellective experiments, evaluating and validating both encoded parameters and

the extensions of VDT. The intellective experiments considered three elements: (1) task

complexity — pooled, sequential, reciprocal and intensive workflow cases —, (2)

organization styles — a Japanese organization style, an American organization style, and

the different levels of matrix strength —, and (3) micro-level behavior patterns — a

Japanese behavior pattern, an American behavior pattern (the defaulted behavior pattern in

VDT), and the possible combinations of the two behavior patterns in a team.

169

In this dissertation, I showed that project managers need to consider not only four

elements — project complexity, organization design, participants’ behaviors, and project

phase — but also linkages among the four elements, in order to lead their project

effectively and efficiently. Based on the direct observations and computational

experiments I draw the following conclusions:

1. There are cultural differences in international joint venture projects. Japanese

and American teams comprising an IJV team have their own sets of cultural

values — i.e., Japanese team members are most likely to show group-based

decision making and communication behaviors, while American team members

are apt to have individual-based decision making and communication behaviors

— and cultural practices — i.e., typical cultural practices of Japanese teams

include a high level of centralization, a high level of formalization and multiple

layers of hierarchy, while typical cultural practices of American teams are low

level of centralization (decentralization), medium level of formalization, and a flat

hierarchy — which combine to form culturally-driven normative systems

2. The Proposed InterCutlural-Virtual Design Team (IC-VDT) model provides an

effective platform for us to understand and analyze cultural impact on project

performance for mixed-cultural teams by extending the current VDT model.

The simulation model incorporated the conception of institutional exceptions as

the consequence of cultural interactions. The model has been validated through

two sets of intellective experiments: one for single cultural cases and one for

mixed cultural cases. The two intellective experiments successfully validated the

reasoning of IC-VDT acquired organizational contingency theory (Thompson,

1967; Galbraith, 1974, 1977) for internal validity, cultural contingency theory

(Hofstede, 1991; Adler,1997), and added refineries — convergence vs.

divergence leadership style (Adler, 1997) and matrix strength (Davis and

Lawrence, 1973)— as propositions to be tested in the research

170

3. Culturally-driven normative systems — cultural values and cultural practices

— do affect, by and large, project performance from an information processing

point of view, and the cultural values and cultural practices have different

impacts on team performance — i.e., changes in behavior patterns had less

impact on team performance than did changes in organization structures. The

impacts of cultural differences increase, as a project becomes more intensive and

more complex. At this stage, the relative contributions of the organization system

or behavior pattern are unknown and cannot be analyzed quantitatively,

qualitative tendencies — i.e., the more flexible American organization structure

enables a bigger impact due to changes in behavior patterns than does the

Japanese organization structure. This is consistent with my limited observations

of US-Japan joint venture projects.

4. Each organization style has its own pros and cons. For instance, the Japanese

organization style shows consistently better project quality risk over all cases, and

better performance (project duration and project cost) in the case of high project

complexity. However, in the case of medium project complexity, the American

organization style shows better performance for cases of medium project

complexity. Additionally, the America organization style is more vulnerable to

the negative impact of low team experience than the Japanese organization

structure.

5. One’s micro-level behavior pattern is positively correlated to one’s organization

structure style. Understanding the qualitative relationships between cultural

practices and cultural values through virtual computational experiments provides

evidence for, and insight into, the evolutionary phenomena of specific

organization structures in each country described by (Greif, 1994)

6. Decentralized organization style such as the American organization style is

preferred for mixed-cultural and culturally diverse teams. Additionally,

drawing on the “matrix swing” concept from Morris (1982), mixed-cultural teams

171

have the potential to create struggles at the beginning or at the end, because

construction projects demand a centralized structure in order to integrate opinions

and judgment, and because high reciprocal interdependence causes a high demand

for information processing.

7. Mixed cultural teams can potentially perform less efficiently — i.e., 70 %

increased project duration — or more efficiently — i.e., 30% improved project

duration — than single cultural teams. My statistical analyses confirmed that

the significance of this hypothesis in terms of project duration and hidden work

volume outcomes was not strong, compared to other hypotheses. However, this

hypothesis was strongly correlated to project quality risks.

8. Matrix strength is a key dimension of management style for global projects and

corporations. Higher matrix strength significantly affected project duration and

hidden work volume, but not project quality risks.

My conclusions are limited by a series of assumptions I made in this dissertation.

I assumed that: team members do not adapt their values or practices during the course of

a given project; participants’ motivation and productivity is consistent during a project;

my dissertation is limited to only two cultural cases (Japanese and American cultures);

and there is no positive innovative impact from cultural interactions.

8.2 Contributions This research draws upon theoretical foundations from several disciplines. The results

and implications of this research contribute to our understanding in the fields of

computational organizational theory, cultural and institutional theory, and comparative

research on Japanese and American corporations.

8.2.1 Contributions to Organization Science

In organization science research, cultural issues have traditionally been dismissed as

esoteric (Martin, 2002). How and why does national culture affect organization

172

structure? Organization research provides surprisingly few answers. The trend toward

globalization has created a demand for research among disciplines as diverse as

organization science, sociology, and anthropology, in order to create a foundation for

new theories about cultural factors. Numerous sociologists have investigated the core

role of cultural values in societies and institutions. In the organization sciences,

“practices” have long been the primary focus of attention in discussions of efficient

organization design (Martin, 2002). The “practice-value” dimensions drawn upon by this

research provide a robust framework for combining cultural and organizational theory.

The simulation approach taken in this research elucidates causal connections

between cultural values and practices at the project-organizational level. The

computational simulation model quantifies the effects of changes in behavior patterns

caused by cultural values, and also the effects of changes in organizational structure types

representing cultural practices. Understanding the relationships between cultural

practices and values through virtual computational experiments clarifies the evolutionary

phenomena of specific organization structures in each country (Greif, 1994). In other

words, normative and regulative institutions in a group can evolve in harmony with the

group’s cultural-cognitive institutions (Scott, 2001). It also pushes the limits of

computational modeling, in particular, the applicability of the VDT model to global

projects. Moreover, this research presents a way to understand and measure the effects of

cultural differences on organization performance.

The practice-value approach also sheds light on the internal complexity within

global project teams — the primary motivation behind developing the “InterCultural

Virtual Design Team” model. IC-VDT addresses internal institutional exceptions

separately from technically driven exceptions. Modeling institutional exceptions as

arising from “practice differences” and multiple decision-making and communication

behavior patterns that are the essence of “value differences” enables the representation of

real global projects, and also allows managers to design better organizations for cross-

cultural teams. This research starts with two cultural cases — American vs. Japanese —

but can ultimately incorporate multi-cultural cases involving additional cultures.

173

8.2.2 Contributions to Cultural and Institutional Theory

Cultural and institutional researchers have conducted investigations at various levels,

ranging from the individual to societal and national levels. One of the principal ways in

which institutional theories differ is in the level at which they are applied (Scott, 2001).

Interestingly, very few researchers have focused on the project-organization level. Issues

arising during international joint-venture projects are a new area of study in the

institutional field.

One of the key contributions of this research to institutional theory is in

addressing the internal complexity of global project teams in terms of both values and

practices. Cultural core values shape the meanings of activities and objectives. Cultural

value differences imply that differences in cultural cognition drive different behavior

patterns. In other words, individuals’ behaviors can be institutionalized by institutional

systems — regulative, normative, and cultural-cognitive systems — in their home

country. Therefore, individuals relatively sustain or persist in their behavior even though

they move to another country.

On the practices side, subgroups of joint-venture project teams have their own sets

of standardized practices, which have been shaped and guided by institutions in their

home countries. These institutionalized practices are the consequence of the different

institutional systems each country possesses. Thus, we can see global project issues as

internal institutional-complexity problems. The internal institutional complexity

discussed in this work gives new insights to cultural and institutional theory.

Additionally, the new model includes the institutional conflicts that emerge from

institutional differences at the project organization level. Recently, there has been an

increased interest in research on institutional conflict and change. However, there are

very few studies about the micro-processes that lead to institutional conflict and its

relation to outcomes, except for studies looking at very long-term outcomes (e.g., Aoki,

2001). Since these are relatively new topics in this field, exploratory studies are called

for to provide descriptive analyses of the processes that trigger institutional conflicts.

174

8.2.3 Contributions to Cross-cultural and Intercultural Research on Japanese and

American Corporations

Many organizational researchers since the 1970s and 1980s have compared and

contrasted national cultures, especially for Japanese vs. American firms. Most of these

studies have been conducted in automobile, utility, and banking companies. Since the

data gathered here relates to the construction industry, this research will contribute to our

understanding of the consistency of Japanese firm behavior across industries and also the

uniqueness of the construction industry.

One of the key benefits of comparative studies is to discover an important

principle by comparing two countries, and then apply the principle to other countries.

Thus, interesting findings from this research regarding correlations between cultural

values and practices can provide a framework to study other cultural cases. The IC-VDT

model has been calibrated and validated by studying these two cultures. However, this

model can be applied to other cultural cases, since many of its parameters are calibrated

based on the national cultural indices proposed by Hofstede, which cover 53 countries

(Hofstede, 1991).

Secondly, many Japanese firm-related studies are from the automobile industry or

the electrical industry. Very few researchers have focused on the construction industry.

My case studies confirmed the empirical findings drawn from previous researchers — i.e.,

my ethnographies support hierarchical structure in Japanese teams as proposed by

Nakane (1970). My case studies confirm the validity of Hofstede’s cultural value

dimensions. Meanwhile, I found different features. For instance, effects of differences in

the Hofstede masculinity-vs.-feminity cultural index were not observed in construction

projects.

Finally, this research sheds new light on institutional exceptions caused by

differing cultural practices. This dissertation captured and modeled only “information

bouncing” phenomena between actors at cross-cultural interfaces. However, simulated

results interestingly indicated that this phenomenon combines with other factors — i.e.,

organization configurations, organization structures, and task workflows — to affect

project performance, either positively or negatively. I believe that this new concept can

stimulate further discussion and arguments in this field. Moreover, intercultural research

175

is a relative new field so that it calls for scholars to theorize, qualify, and quantify the

concept of institutional exceptions.

8.2.4 Contribution to Practice

Do managers see culture? Managers seem to demonstrate a cultural blindness (Adler,

1983; Anderson, 1966). In many instances people associate recognizing cultural

differences with simplistic, primitive, and/or immoral thinking (Hofstede, 1991; Adler,

1997). Social norms, in particular in the “melting pot” of North America, often

encourage managers to blind themselves to gender, race, and ethnicity — i.e., people

might neither see nor want to see differences. This approach potentially causes problems

because it can affect their judgment. For instance, my research found that people from

different cultural groups behave differently, and then these differences affect project

outcomes. People from one cultural group are not inherently any better or worse than

those from other groups: they are simply different. I emphasize here that it is important

neither to ignore nor to overreact to cultural differences. Choosing not to see cultural

diversity may limit our ability to manage projects. Judging cultural differences as good

or bad can lead to inappropriate, offensive, and unproductive racist and ethnocentric

attitudes and behaviors. Reasons of high failure rate in IJV projects may be that IJV team

members might fall into either of cases — ignoring or overreacting to cultural differences

—, leading to misjudgment.

How can we explicitly know and realize what are cultural differences in a global

project? How can we know and understand what are sequences of cultural differences?

How can project managers successfully lead a project? This research is an initial step to

answer these questions using a computational simulation model. A computational

simulation model such as IC-VDT discussed in this dissertation can provide an objective

criterion about which practice style a project team selects at which project stage and in

which project complexity.

This research begins with the bi-cultural case of Japanese and American cultures.

The findings of this research suggest the potential to use IC-VDT to model other cultural

cases. Successfully modeling other cultural cases will validate IC-VDT’s robustness and

applicability to global projects that involve participants from dozens of countries.

176

The research also contributes to provide a managerial tool for project managers to

assess project organizational risks on global projects. A manager can use this tool to

assess what the risks are and explore how to mitigate coordination and institutional risks

by changing organization styles.

Finally, the research efforts can provide a training tool for participants in building

cohesive global teams. In the early phases when the project is being shaped, project

participants can understand, through simulation activities, other partners’ approaches and

the reasoning behind their behaviors. In particular, virtual experiences become important

when a project team needs to face a new or unknown situation (Levitt, 2004).

8.3 Future Research This research focuses on project organizations in global projects. As shown in Figure 8.1,

this research extends VDT research to address cross-cultural cases by adding different

behavior patterns and different practices, which generate internal complexity. Looking at

internal complexity in global project teams help us understand and analyze the challenges

of coordination among people and groups. The cultural value-practice dimension is a

useful framework to generalize and encode internal complexity issues into simulation

models. However, the IC-VDT model retains the core information processing abstraction

of VDT and so does not take into consideration the emotional and reactive behaviors of

individual actors when they encounter different cultural values and practices. Some

emotionally charged and reactive behaviors of individual actors include trust vs. distrust,

and negative vs. positive aspirations and motivations. Modeling trust relationships

among people and groups is an interesting topic for future global project-organization

studies. Further exploratory investigation and ethnographic data are needed to analyze

the reactive behaviors of individual actors.

Additionally, this research does not include the external environmental pressures

that project teams face, nor the kinds of learning that can take place between two cultures.

It would be interesting to capture the external institutional complexity surrounding global

project teams, because this research only sheds light on internal complexity. Regulative

institutional elements (Scott, 2001) are, needless to say, also prime foci. Finally, how

actors learn cultural values and practices from each other remains interesting and largely

177

unexplored. Therefore, future research might include the addition of reactive behaviors

of individual actors, external institutional complexity, and non-linear learning effects.

Figure 8.1 Anticipated Contributions to VDT Research

behaviors (trust, conflicts, innovation ideas)

Actors IC-VDT The current VDT model

The IC-VDT model Multiple patterns of

micro-level behaviors

One pattern of micro-level behaviors

No learning Technically-driven complexity

Culturally-driven complexity

Complexity Learning

Future Research Actors

The current VDT model

The IC-VDT model Multiple patterns of micro-level behaviors

Future Research

One pattern of micro-level behaviors

No learning Technically-driven complexity

Non linear learning Culturally-driven complexity

Learning Complexity Institutional complexity

This figure illustrates the research trajectory of the Collaboratory for Research on Global Projects (CRGP) and the Virtual Design Team (VDT) group. My contributions are to address internal complexity and different micro-level behavior patterns as the key cultural phenomena arising in international joint-venture projects. Future research should look into reactive behavior on the actor axis, external complexity along the complexity axis, and non-linear learning issues on the time axis.

178

Appendix A: American Behavior Pattern

(Defaulted Version of Micro-Level Behavior)

1) Decision Making Policy

High Medium Low PM 0.6 0.2 0.1 SL 0.3 0.6 0.3 ST 0.1 0.2 0.6 DH 0.8 0.5 0.2 DM 0.2 0.5 0.8

Note: The Decision-Making-Policy matrix determines which role makes decisions about handling exceptions, based on the project’s level of centralization. The matrixvalues show the probability that a particular role will make the decision when the centralization is set at high, medium, and low.

179

should wait for a decision regarding an exception to be made before making the decision themselves. The length of time is measured in minutes.

2) Type of Decision (rework, correct, or ignore)

Rework Correct Ignore PM 0.65 0.30 0.05 SL 0.40 0.40 0.20 ST 0.05 0.35 0.60 DH 0.65 0.30 0.05 DM 0.05 0.35 0.60

3) Tolerance in Waiting for Decisions

PM 480 SL 960 ST 960 DH 480 DM 960

Note: The Type-of-Decision matrix determines how an exception should be dealt with, based on project's centralization policy. The assumption is that higher-level roles (for example, project managers) tend to make more rework decisions.

Note: The Tolerance-in-Waiting-for-Decision matrix determines how long each role

4) Attendance To Communication

Communication 0.5 1 2 Meet 0.7 1 1

Formalization High Medium Low

Noise 1 1 1 Note: The Attendance-To-C

5) Response Probability of Communication

Matrix High Medium Low Communication 0.9 0.7 0.6

Meet 0.6 0.7 0.9 Noise 1 1 1

Note: The Response-Probability-Of-Communi

matrix strength of the organization. Project organizfunctional teams have low matrix strength. The ma

6) Demand Volume for Communication.

nform

a change co municat , meetinrix S setting h, medium, or low.

ommunication matrix helps to determine the probability of each kind of information exchange taking place — information exchange, meetings, or noise — depending on the level of project formalization.

cation matrix determines the probability that a position will attend to a given type of information exchange, depending on the

ations have high matrix strength; trix rows show the type of

inform tion ex m ion gs, and noise and the columns show the Mat trength , hig

Decision Exception Exchange Meeting Noise PM 8 28 28 0 10 SL 8 58 28 0 10

I ation

ST 8 58 28 0 10 DH 8 28 28 0 10 DM 8 58 28 0 10

Not e Dema olume-For-Communication matrix defines the volume in minutesof e ype of m e for each role in an organization. The matrix rows show the role d the co essage types.

e: Th nd-Vach t essags, an lumns show the m

180

Appendix B: Japanese Behavior Pattern

(Adjusted Version of Micro-Level Behavior)

1) Decision making policy

High Medium Low PM 0.7 0.25 0.15 SL 0.2 0.55 0.35 ST 0.1 0.2 0.5 DH 0.85 0.55 0.25 DM

Note: The Decision- ing-Po atrix ines which role makes decisions abou ing excep s, base the pro s level of centralization. The matrix values show the proba lity that a particular role will make the decision when the

Mak licy m determt handl tion d on ject’

bi

0.15 0.45 0.75

2) Type of decision (rework, correct, or ignore)

Rework Correct Ignore PM 0.55 0.42 0.03

centralization is set at high, medium, and low.

SL 0.41 0.45 0.14 ST 0.26 0.33 0.41 DH 0.70 0.27 0.03 DM 0.07 0.41 0.52

3) Tolerance in Waiting for Decisions

PM 576 SL 1152 ST 1152 DH 576 DM 1152

Note: The Typ s de w tiobased on project's ntralizat olicy. T ssumptio s that hi level roles (for exam , project m make more rework decisions.

e-of-Deci ion matrix termines ho an excep n should be dealt with, ce ion p he a n i gher-

ple anagers) tend to

Note: The Tolerance-in-Waiting-for-Decisions matrix determines how long each role should wait for a decision regarding an exception to be made before making the decision themselves. The length of time is measured in minutes.

181

4) Attendance To Communication

Communication 0.7 1.2 2.2 Meet 0.8 1.1 1.0

Formalization High Medium Low

Noise 1.0 1.0 1.0

Note: The Attendance-To-C

ommunication matrix helps to determine the probability ofeach kind of information exchange taking place — information exchange, meetings, or noise — depending on the level of project formalization.

5) Response Probability of Communication

Matrix High Medium Low Communication 1.2 0.9 0.8

Meet 0.8 0.9 1.2 Noise 1.0 1.0 1.0

Note: The Response-Probability-Of-Communi

matrix strength of the organization. Project organifunctional teams have low matrix strength. The ma

a change co municat , meetinrix S setting h, medium, or low.

cation matrix determines the probabilitythat a position will attend to a given type of information exchange, depending on the

zations have high matrix strength; trix rows show the type of

inform tion ex m ion gs, and noise and the columns show the Mat trength , hig

6) Demand Volume for Communication.

nform Decision Exception Exchange Meeting Noise PM 15 35 35 0 10 SL 15 65 35 0 10

I ation

ST 15 65 35 0 10 DH 15 35 35 0 10 DM 15 65 35 0 10

e

Not e Dema olume-For-Communication matrix defines the volume in minutesof e ype of m e for each role in an organization. The matrix rows show the role d the co show the message types.

e: Th nd-Vach t essags, an lumns

h

182

183

Abegg C., (1958 e Japa Facto spects of Its Social Organization, New

Free Pre 958, pp

A

97

Adler, N.J., (1983), “Cross-Cultural Management: Issues to be Faced,” International

Studies of Management and Organization, vol.13, no.1-2 (spring-Summer 1983),

pp7-45

Aoki, M., (1992), “Decentralization-Centralization in Japanese Organization: A Duality

ple,” Jap e Polit Econom ol. 3, Stanford University Press, 1992,

-169

ald Dore,

4

ki, M., (2001), Toward A Comparative Institutional Analysis, MA: MIT Press, 1992,

3rd ed., 2001

ember Attitudes and Task Performance of

tercult s s S cho ol.69, 1966, pp

-319

Balig ., and Bu , R., (19 , “Descri and Des ing Org ational Structure

Process ternatio Journal olicy Analysis Information Systems, Vol.

981, pp -266.

Sciences, Vol. 1 (2), May 1980, pp. 133-165.

aum, J., (2002), The Blackwell Companion to Organizations, Malden, MA: Blackwell

publisher, 2002

REFERENCES

len, J. ), Th nese ry: A

York: ss, 1 129

dler, N.J., (1997), International Dimensions of Organizational Behavior, Cincinnati,

Ohio, South-Western College Publishing, 3rd ed., 19

Princi anes ical y, V

pp 142

Aoki, M., (1988), Information, Incentives, and Bargaining in the Japanese Economy,

Cambridge, UK and New York: Cambridge University Press, 1988

oki, M., (1994), Japanese Firm: It’s Competitive Sources, co-edited with Ron

Oxford, UK: Oxford University Press, 199

o

A

A

Anderson, L.R., (1966), “Leader Behavior, M

In ural Discu sion Group ,” Journal of ocial Psy logy, V

305

h, H rton 81) bing ign aniz

and ,” In nal of P

5, 1 . 251

Baligh, H., and Damon, W., (1980), "Foundation for a Systematic Process of

Organization Structure Design," Journal of Information and Optimization

B

183

Beamish, P. W., (1985), “Characteristics of Joint Ventures in Developed and Developing

Countries,” Columbia Journal of World Business, Fall: 13-19, 1985

eamish, P. W., and Delios, A., (1997), Incidence and Propensity of Alliance Formation,

ol. 27(1), 2002, pp. 14-

Botti,

on”, Journal of Organization, Vol. 2, 1995, pp. 55-86

al and Mathematical Organization Theory, Vol. 1, 1995, pp. 57-71

ement Science, Vol. 48-11, 2002,

Burton, R. M., (2003), “Computational Laboratories for Organization Science: Questions,

l.

Buckley, P. J., and Casson, M., (1988), “A Theory of Cooperation in International

Carley ting Computational Models, Carnegie, PA: Carnegie

Carley

d Mathematical Organization

B

in Cooperative Strategies: European Perspectives, San Francisco, CA: New

Lexington Press, 1997

Bells and Kozlowski (2002), “A Typology of Virtual teams: Implications for Effective

Leadership,” Group and Organization Management, V

49

H.‚ (1995), “Misunderstandings: A Japanese transplant in Italy Strives for Lean

Producti

Burton, R. M, and Obel, B., (1995), “The Validity of Computational Models in

Organization Science: From Model Realism to Purpose of the Model,”

Computation

Burton, R. M., and Obel, B., (2004), Strategic Organizational Diagnosis and Design,

Norwell, MA: Kluwer academic publisher, 1998, 3rd ed., 2004

Burton, R. M., Lauridsen, J., and Obel, B., (2002), “Return on Assets Loss from

Situational and Contingency Misfits,” Manag

pp.1461-1485.

Validity and Docking,” Computational & Mathematical Organization Theory, Vo

9, 2003, pp. 91-108

Business,” In Farok Contractor and Peter Lorange, Editors, Cooperative strategies

in International business, 31-53, Lexington, Mass: Lexington Books

, K. M., (1996), Valida

Mellon University, 1996

, K. M., (1995), “Computational and Mathematical Organization Theory:

Perspective and Directions,” Computational an

Theory, Vol. 1(1), 1995, pp.39-56

184

Caroll, E., and Burton, R.M., (2000), “Organizations and Complexity: Searching for the

Edge of Chaos,” Computational and Mathematical Organization Theory, Vol. 6,

Cavus

An Empirical Investigation”, J, Business Research, Vol. 12,

Chacher

Integrated Concurrent Engineering: Grounded Theoretical Factors and Risk

Christi rdination in

ring,

Clough

Cole, R American,

1989

Cohen

ring in Project Teams, PhD. Dissertation, Department of Civil

Contra

for cooperative ventures,” Cooperative strategies in

Cox, T

ol.5, 1991,

Cullen, J.B., Johnson, J. L., and Sakano, T., (1995), “Japanese and Local Partner

stments

o.1,

Cyert, R. and March, J. (1963), A behavioral Theory of the firm, Englewood Cliffs, NJ:

Prentice-Hall, 1963

2000, pp. 319- pp.337

gil, T. S. and U. Yavas, (1984), “Transfer of Management Know-how to

Developing Countries:

1984, pp. 35-50

e, J., John K., and Levitt, R., (2002), Observation, Theory, and Simulation of

Analysis Using Formal Models, Stanford CIFE Technical report, 2002

ansen, T., (1993), Modeling Efficiency and Effectiveness of Coo

Engineering Design Teams: VDT-the Virtual Design Team, Civil Enginee

Stanford, CA: Stanford University, 1993

, R. C., Sears, G., and Sears, S., (2000), Construction Project Management, New

York, NY: John Wiley & Sons Inc., 4th ed, 2000

. E., (1989), Strategies for learning: Small Group Activities in

Japanese, and Swedish Industry, Berkeley, University of California Press,

, G.P., (1992), The Virtual Design Team: An Object-Oriented Model of

Information Sha

Engineering, Stanford, CA: Stanford University, 1992

ctor, F. J., and Lorange, P., (1988), “Why should firms cooperate? The strategy

and economics basis

international business, 3-28, Lexington, MS: Lexington Books

.H., and Blake S., (1991), Managing Cultural Diversity: Implications for

Organizational Competitiveness, Academy of Management Exective, V

No. 3

Commitment to IJVs: Psychological Consequences of Outcomes and Inve

in the IJV Relationship,” Journal of International Business Studies, Vol. 26, N

1995, pp.91-115

185

Davis, S., M., (1984), Managing Corporate Culture, Cambridge, MA: Ballinger, 1984

S., and Lawrence, P., (1977), Matrix, Addison Wesley, 1977

gio, P. J., and Powell, W

Davis,

DiMag . W., (1983), “Iron Cage Revised: Institutional

ican

Dooley, K., (2002), “Simulation Research Method,” The Blackwell Companion to

Eisenh f

ed on

b-

Isomorphism and Collective Rationality in Organizational Fields,” Amer

Sociological Review 48:147-60, 1983

Organizations, Chapter 36, Malden, MA: Blackwell, 2004

ardt, K.M., (1989), “Building Theories from Case Study Research,” Academy o

Management Review, 532-50, 1989

Engineering News Record (ENR) (2004), Top 225 international contractors, access

November 7th 2004: <http://enr.construction.com/features/international/archives/040823

1.asp>

Etzioni, A., (1964), Modern Organizations, Englewood Cliffs, NJ: Printed Hall, 1964

an, D.C. and Arnold, H. J., (1983), MaFeldm vidual and Group Behavior in

Flanag

national

Fridsm rotocols for Organizational

Frucht d multi-modal

journal

Fujimoto, T., (2004), Capability-Building Competition: Why is the Japanese Automobile

okyo,

Galbra

1974

naging Indi

Organizations, New York, NY: McGraw-Hill, 1983

an, R., (1994), “The Features of Successful Construction Companies in the

International Construction Market,” proceedings of the A.T. Elkin Inter

Seminar on strategic planning in construction companies, 1994, pp 304-318

a, D.B., and Thomsen, J., (1998), “Modeling Medical P

Simulation: An Information-Processing Approach,” Computational and

Mathematical Organization Theory, Vol. 4, 1998, pp71-95

er, R., and Townsend, A., (2003), “Multi-cultural dimensions an

communication in distributed cross-disciplinary teamwork,” International

of engineering education IJEE 2003, vol19, nr.1, 53-61

Industry Strong (能力構築競争:日本の自動車産業はなぜ強いのか),” T

JP: Chuo-Kouron Sin-sya, 2004

ith, J., (1974), Organization design: An Information Processing View,

INTERFACES 4,

Galbraith, J., (1977), Organization Design, Addison-Wesley, New York, NY, 1977

186

Glaser, B.M., and Strauss, A.L., (1967), The Discovery of Grounded Theory: Strategies

967

unication

Graham rgaining: Doing Business with the

Granov he Problem of

Greif, rganization of Society: A Historical and

Hart, W 998), “What is Intercultural Relations?” The Edge: The E-Journal of

Hawle

Heston ity comparisons:

Hofste

Hong, e Performance of Contractors in Japan, the UK

uction

mics, Vol20, 2002, pp425-435

Influences on Team Performance through Virtual Experiments," Collaboratory

for Qualitative Research, Hawthorne, NY: Aldine de Gruter, 1

Glenn, E. S., and Glenn, C. G., (1981) Man and Mankind: Conflict and Comm

Between Culture, Norwood, NJ: Ablex Pub Corp., 1981

, J. L., and Sano, Y., (1984), Smart Ba

Japanese, Cambridge, MA: Ballinger Publish Inc., 1984

etter, M., (1985), "Economic Action and Social Structure: T

Embeddedness," The American Journal of Sociology, Vol .91, 1985, pp481-510

A., (1994), “Cultural Beliefs and the O

Theoritical Reflection on Collectivist and Individualist Societies,” The Journal of

Political Economy, Vol. 102, Issue 5, Oct., 1994, pp912-950

.B., (1

Intercultural Relations, summer, Vol. 1 (3), 1998

Hannan, M. T., and Freeman, J., (1984), “Structural Interia and Organizational Change,”

American Sociological Review, 49:149-64, 1984

Hannan, M. T., and Freeman, J., (1989), Organizational Ecology, Cambridge, MA:

Harvard University press, 1989

y, A., (1950), Human Ecology, New York: Ronald Press, 1950

Hawley, A., (1968), Human Ecology: A Theoretical Essay, Chicago: University of

Chicago press, 1968

, A., and Summers, R., (1996), “International price and qual

potential and pitfalls,” American Economic review, Vol.86, 20-4, 1996

de, G.,(1980), Cultures’s Consequences: International Differences in Work

Related Values, Bevery Hills, Sage Inc., 1980

Hofstede, G., (1991), Cultures and Organizations: Software of the Mind, Intercultural

Cooperation and its Importance for Survival, New York, NY: McGraw-Hill, 1991

X., and Proverbs, D., (2002), “Th

and the USA: A Comparative Evaluation of Construction Cost,” Constr

Management and Econo

Horii, T., (2003), "Cross-Cultural Teams: Modeling and Qualitative Analysis of

187

Research for Global Projects (CRGP) Technical Report #001, Stanford Universit

Stanford CA, 2003

y,

al and Mathematical

Horii, ing Cultural Influences

CSOS

Horii,

tion Research

Horii, rmative Systems on

nt Systems (NorMAS2005) as a

ntion,

sfield, England, April 12-15, 2005

., (2004), Culture,

4

3, pp883-

Howel ed., Pacific Grove, CA, 2002

Jeremi n of Organizational

Horii., T., Jin, Y., and Levitt. R.E., (2005), “Modeling and Analyzing Cultural

Differences on Project Team Performance,” Computation

Organization Theory, Vol.10, No.4, January 2005, p305-321

T., Jin, Y., and Levitt, R.E. (2004), "Modeling and Analyz

on Team Performance through Virtual Experiments," Proceeding of the NAA

Conference 2004, Pittsburgh, PA (The best paper award at the PhD Abstract

Competition), 2004

T., Levitt, R.E., and Jin, Y. (2005), "Cross-Cultural Virtual Design Team: Cultural

Influence on Team Performance of Global Projects," Construc

Congress (CRC) 2005, San Diego, CA, April 5-7, 2005

T., Y. Jin, and R.E. Levitt, “Impact of Multiple No

Organization Performance of International Joint Venture Projects”, 1st

International Symposium on Normative Multi-Age

part of the 2005 Artificial Intelligence and Simulation Behavior (AISB) conve

University of Hertfordshire, Hart

House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., and Gupta, V

Leadership, and Organizations-the GLOBE study of 62 Societies, Sage

Publications, Inc., 200

Howard, A., Shudo, K., and Umeshima, M., (1983), “Motivation and Values Among

Japanese and American Managers,” Personnel Psychology, Vol. 36, 198

898

l, D. C., Statistical Methods for Psychology, 5th

Ibrahim, R., and Nissen, M.E., (2004), "Simulating Environment Contingencies Using

SimVision," Proceedings of the North American Association of Computational

Social and Organization Science Conference, Carnegie-Mellon University,

Pittsburgh, Pennsylvania, 27-29 June, 2004

ah, J., Sullivan, and Nonaka, I., (1986), “The Applicatio

Learning Theory to Japanese and American Management,” Journal of

International Business Studies, Vol. 17, No3, autumn, 1986, pp127-147

188

Jin, Y.

s,” Computational and Mathematical Organization Theory,

KHosr

r Computational

KHosr ation

amming,” In Keijzer, Maarten (editor), Late-Breaking Papers

; July,

ntary

2001,

Koike,

demy of

l. 45, No.1, 2002, pp. 215-233.

Kunz, J. C., Christiansen, T. R., Cohen, G. P., Jin, Y., and Levitt, R. E., (1998), “The

Lachm

al Framework,” Management Science,

Lauren gement,”

and Levitt, R.E., (1996), “The Virtual Design Team: A Computational Model of

Project Organization

Vol. 2(3), 1996, pp171-196

aviani, B., and Levitt, R.E., (2004), “Organization Design Using Genetic

Programming,” Proceedings of North American Association fo

Social and Organizational Science (NAACSOS) Conference, Pittsburgh, PA, June

27-29, 2004

aviani, B., Levitt, R.E., and Koza, J.R., (2004), “Organization Design Optimiz

Using Genetic Progr

at the 2004 Genetic and Evolutionary Computation Conference (GECCO-2004),

International Society of Genetic and Evolutionary Computation, Seattle, WA

2004

Klahr, D., and Simon, A. H., (2001), Studies of Science Discovery: Compleme

Approaches and Convergent Findings, Psychological Bulletin, 125 (5),

pp524-543

K, (1988), Understanding the Industrial Relations in Japan, London: Macmillan,

1988

Kostova, T., and Roth, K., (2002), “Adoption of An Organization Practice by Subsidiaries

of Multinational Corporations: Institutional and Relational Effects,” Aca

Management Journal, Vo

Kravis, I. B., (1984), Comparative Studies of National Incomes and Prices, Journal of

economic literature, 22, 1-39

Virtual Design Team: A Computational Simulation Model of Project

Organizations,” Communications of the Association for Computing Machinery

(CACM), Vol. 41(11), 1998,

an, R., Nedd, A., and Hingings, B., (1994) “Analyzing Cross-National

Management and Organizations: A Theoretic

Vol. 40, No#1, January, 1994, pp. 40-55

t, A., (1983), “The Cultural Diversity of Western Conceptions of Mana

International Studies of Management and Organization, Vol. 13, no.1-2

189

Levitt, R. E., and Kunz, J C., (2002), Design Your Project Organization as Engineers

Design Bridges, CIFE technical paper, Stanford University, September, 2002

R. E., Cohen, G. P., Kunz, J.Levitt, C., Nass, C.I., Christiansen, T., and Jin, Y., (1994),

.

: Lawrence Erlbaum Associates Publishers, 1994

Levitt, Fyall, M., Bjornsson, P., Hewlett, W., III, and Casebolt, T., (2002), “When

ling in the Social Sciences, UCLA Lake Arrowhead Conference

Levitt, orii, T., Mahalingam, A., Orr, R., and Taylor, J., “Understanding and

dy

an,

Louis, M., (1985), “An Investigator’s Guide to Workplace Culture,” in P. Forest, L.

ure, pp. 73-

McFar (1933), Twenty-First-Century

“The Virtual Design Team: Simulating How Organization Structure and

Information Processing Tools Affect Team Performance,” in Carley, K. M. and M

J. Prietula, editors, Computational and Mathematical Organization Theory,

Hillsdale, NJ

Levitt, R. E., Thomsen, J., Christiansen, T. R., Kunz, J.C., Jin, Y., and Nass, C., (1999),

“Simulating Project Work Processes and Organizations: Toward a Micro-

Contingency Theory of Organizational Design,” Management Science, Vol. 45

(11), November, 1999, pp1479-1495

R. E.,

Information Flow in Project Organizations Becomes Turbulent: Toward an

Organizational Reynolds Number,” Proceedings of the Conference on Agent-

Based-Mode

Center, May 2002, pp3-7

R.E., H

Managing the Effects of Institutional Differences in Global Projects,” ASCE

Specialty Conference on Management and Leadership in Construction, Hilton

Head, South Carolina, USA, March, 2004

Levy, S.M., (1990), Japanese construction: An American Perspective, Van stand

Reinhold, New York, 1990

Lincoln, J. R., and Kalleberg, A. L., (1990), Culture, Control, and Commitment: A Stu

of Work Organization and Work Attitudes in the United States and Jap

Cambridge, UK: Cambridge University Press, 1990

Moore, M. Louis, C. Lundberg, & J. Martin (Eds), Organizational Cult

94, Beverly Hills, CA: Sage publication Inc., 1985

land, L. J., Senen, S., and Childress, J. R.,

Leadership, New York, Leadership Press, 1933

190

Mahalingam, A. and Levitt, R.E., (2004), “Challenges on Global Projects — An

Institutional Perspective,” Proceedings of the International Symposium of the CIB

Mahal

of the

ent Systems: The

Malon

p.1317-1332

Martin :

Maslo f Human Motivation,” Psychological Review, July, 1943,

Mintzb n

Nakan d LA, CA: University of California

Nasrallah, W. F., and Levitt, R. E., (2001), “An Interaction Value Perspective on Firms

Nasral opularity in

tion

Nissen eling of Knowledge

4, pp169-

W92 on Procurement Systems: Project Procurement for Infrastructure

Construction, Chennai, India, 7-10 January, 2004

ingam, A., Levitt, R.E., and Scott, W.R., (2005), “Cultural Clashes in International

Infrastructure Development Projects: Which Cultures Matter?” Proceedings

International Symposium of CIB W92/TG23/W107 on Procurem

impact of Cultural Differences and Systems on Construction Performance, Las

Vegas, USA, 8-10 February, 2005

e, T., (1987), “Modeling Coordination in Organizations and Markets”,

Management Science, Vol. 33, 1987, p

March, J. G. and Simon, H. (1958), The Behavioral Theory of the Firm, Englewood

Cliffs, NJ: Prentice –Hall, 1958

, J., (2002), Organizational Culture; Mapping the Terrain, Thousand Oaks, CA

Sage publication Inc., 2002

w, H, (1943), “A Theory o

pp. 370-396

erg, H., (1980), “Structure in 5’s: A Synthesis of the Research on Organizatio

Design,” Management Science, Vol. 26-3, 1980, pp.322-341.

e, C., (1970), Japanese Society, Berkeley an

Press, 1970

of Differing Size,” Computational and Mathematical Organization Theory, Vol.

7(2), 2001, pp.113-144

lah, W. F., Levitt, R. E., and Glynn, P., (1998), “Diversity and P

Organizations and Communities,” Computational and Mathematical Organiza

Theory, Vol. 4(4), 1998, pp.347-372.

, M.E., and Levitt, R.E., (2004), “Agent-Based Mod

Dynamics,” Knowledge Management Research and Practice, Vol. 2, 200

183

191

Nonaka, I., and Takeuchi, H., (1995), The Knowledge-Creating Company, New Y

NY: Oxford University Press, 1995

ly, A., Charles, and Pfeffer, J., (2000), “Hidden Value: How Great Com

Achieve Extraordinary Results With Ordinary Peop

ork,

O’Reil panies

le,” Harvard Business School

Orr, R

e Benefits, Bangkok, Thailand, 17-19

Orr, R ents: A knowledge-based

B

t Systems: The impact of Cultural Differences

.

Ouchi, w American Business can Meet the Japanese

Rhines 993), Globalization: Six Keys to Success in a Changing World, The

Robins egotiation and

Ronen, S., and Shenkar, O., (1985), “Clustering Countries on Attitudinal Dimensions: A

Sathe, elated Corporate Realities: Text, Cases, and Readings

Irwin,

Schein d Leadership, San Francisco, CA: Jossey-

Schnei

gement,” Human Resource Management, vol. 27, no.2

Press, 2000

., (2004), “Coping with cognitive-cultural, normative and regulative institutional

asymmetry on global projects: A learning perspective,” Proceedings of the

International Symposium of the CIB W107 on Globalization and Construction:

Meeting the Challenges, Reaping th

November, 2004

., (2005), “Strategies to succeed in foreign environm

contingency approach,” Proceedings of the International Symposium of the CI

W92/TG23/W107 on Procuremen

and Systems on Construction Performance, Las Vegas, NV, 8-10 February, 2005

W. G., (1981), Theory Z: Ho

Challenge, Reading, MA: Addison-Welsley, 1981

mith, S. H., (1

American Soceity for Training and Development, Richard D. Irwin, Inc., 1993

on, R., (1997), “Errors in Social Judgement: Implications for N

Conflict Resolution,” Class notes in Harvard, Harvard University, 1997

Review and Synthesis,” Academy of Management Review, Vol. 10, 1985, pp.435-

454

V., (1985), Culture and R

on Organizational Entry, Establishment, and Change, Homewood, IL: R.D.

1985

, E., (1985), Organizational Culture an

Bass, 1985, 2nd ed., 1992

der, S., (1988), “National vs. Corporate Culture: Implications for Human Resource

Mana

192

Schwartz, H., and Davis, S. (1981), “Matching Corporate Culture and Business Strateg

Organizational Dynamics, 1981, pp.

y,”

30-48

Scott, en Christensen, (1995), The institutional Construction of

,

Sengo ed

Simon ministrative Behavior: A Study of Decision-Making Process in

Spradl

Tatum

Taylor sity,” Academy of Management

Taylor

Thomsen, J., (1998), Virtual Team Alliance (VDA): Modeling the Effects of Goal

Thomsen, J., Levitt, R. E., Kunz, J., Nass, C., and Fridsma, D., (1999), “A Trajectory for

zation Theory, Vol. 5(4), 1999, pp.385-401

ation, Vol. 12, 1982,

pp.139-169

Scott, W. R., (2001), Institutions and Organizations, Thousand Oaks, CA: Sage

publications, Inc., 2nd ed., 2001

W.R., and Sor

Organizations; International and Longitudinal Studies, Thousand Oaks, CA, Sage

1995

ku, T., (1985), Willing Workers: the Work Ethics in Japan, England, and the Unit

States, Westport, CT: Quorum Books, 1985

, H. A., (1945), Ad

Administrative Organization, New York: Free Press, 1945 (4th ed., 1997)

ey, P. J., (1979), The Ethnographic Interview, New York, NY: Holt, Rinehart and

Winston Inc., 1979

, C.B., (1983), Decision-Making in Structuring Construction Project

Organizations, Civil Engineering, Stanford, CA: Stanford University, 1983

C. and Stacy B., (1991), “Managing Cultural Diver

Executive, Vol. 5, 1991, pp.45-55

, J. E., and Levitt, R.E., (2005), “Inter-Organizational Knowledge Flow and

Innovation Diffusion in project Based Industries,” Proceeding of the Hawaii

International Conference on Systems Science 38, Hawaii, USA, 2005

Thompson, J. D., (1967), Organization in Action: Social Science Bases in Administrative

Theory, New York, NY: McGraw-Hill, 1967

Incongruency in Semi-Routine, Fast-paced Project Organizations, Civil

Engineering, Stanford, CA: Stanford University, 1998

Validating Computational Emulation Models of Organizations,” Computational

and Mathematical Organi

Triandis, H., C., (1982), “Dimensions of Cultural Varian as Parameter of Organizational

Theories,” International Studies of Management and Organiz

193

Trompenaar, A., (2004), Managing People Across Cultures, Capstone, Ltd., paperback,

2004

Ventrinsky, I. K. Tse, Wehrung, D.A., and Lee, K., (1990), “Organizational Desi

Management Norm

gn and

ative Study of Managers Perception in the

gement,

Weber, M., (1924), The Theory of Social and Economic Organization, translated and

924

Wong, ., (2000), “Virtual Teams: What are Their Characteristics,

l

tion

.G.

and

allinger, 1988

2

s: A Compar

People’s Republic of China, Hong Kong and Canada,” Journal of Mana

Vol. 16, 1990, pp.853-867.

edited by A. M. Henderson and Talcott Parsons, New York, NY: Oxford

University Press, 1924 (1947)

Weber, M., (1924), Economy and Society: An Interpretive Sociology, edited by

Guenenther Roth and Claus Wittich, New York, NY: Bedminister press, 1

(1968)

S.S and Burton, R. M

and Impact on Team Performance?” Computational and Mathematica

Organization Theory, Vol. 6, 2000, pp. 339-360

Yin, R. K., (1984), Case Study Research: Design and Methods, Sage Publications,

Thousand Oaks, CA, 1984, 3rd edi

Ziller, R. C., (1972), “Homogeneity and Heterogeneity of Group Membership,” in C

McClintock, ed., Experimental Social Psychology, New York: Holt, Rinehart

Winston, 1972, pp 385-411

Zucker, L. G., (1988), Institutional Patterns and Organizations: Culture and Environment,

edited by Lynne G. Zucker, Cambridge, MA: B

Zolin, R., (2002), Trust in Cross-Functional, Global teams: Developing and

Validating a Model of Inter-Personal Trust in Cross-Functional, Global

Teams, Stanford University, Doctoral Dissertation, 200

194