the organizational performance of · 2013. 10. 17. · behavior patterns in a project. ic-vdt also...
TRANSCRIPT
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” CaseIntellective 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 WorkflowNote: 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-Communication 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-Communimatrix 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