planned change in organization style based on the laboratory approach

16
Planned Change in Organization Style Based on the Laboratory Approach Author(s): Robert T. Golembiewski and Stokes B. Carrigan Source: Administrative Science Quarterly, Vol. 15, No. 1 (Mar., 1970), pp. 79-93 Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management, Cornell University Stable URL: http://www.jstor.org/stable/2391191 . Accessed: 12/06/2014 12:54 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaborating with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly. http://www.jstor.org This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PM All use subject to JSTOR Terms and Conditions

Upload: robert-t-golembiewski-and-stokes-b-carrigan

Post on 15-Jan-2017

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Planned Change in Organization Style Based on the Laboratory Approach

Planned Change in Organization Style Based on the Laboratory ApproachAuthor(s): Robert T. Golembiewski and Stokes B. CarriganSource: Administrative Science Quarterly, Vol. 15, No. 1 (Mar., 1970), pp. 79-93Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management,Cornell UniversityStable URL: http://www.jstor.org/stable/2391191 .

Accessed: 12/06/2014 12:54

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaboratingwith JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.

http://www.jstor.org

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 2: Planned Change in Organization Style Based on the Laboratory Approach

Robert T. Golembiewski and Stokes B. Carrigan

Planned Change in Organization Style Based on the Laboratory Approach

This study reports the design and results of an effort to change the organization style of a sales unit in a business organization. The learning design was derived from the laboratory approach to organization development, and sought to create a specific kind of social order as well as to provide experience with appropriate skills and attitudes. Changes in organization style were measured with Likert's profile of organizational characteristics. A one-week learning experience helped induce significant changes in self-reports by managers about the style of inter- personal and intergroup relations in their organization, judging from before and after administrations of the profile. The bulk of the learning time was spent in a sensitivity training session, which was intended to prepare subordinates for a confrontation with their superiors concerning the needs of both as they were variously met by their unit's interpersonal and intergroup climate. The entire man- agerial population was exposed to the learning design, so that there was no con- trol group. Therefore the changes in self-reports can only be tentatively attributed to the experimental design, rather than to random factors or the passage of time.

This article reports the design and results of a planned change in the style of a unit of a large organization, using a learning design intended to change patterns of interpersonal and intergroup relations. The specific target was a team of 16 managers who supervised the national sales of a well-known, personal- consumption product. The men represented three hierarchical levels, and had 80 sales- men in the field reporting to them. Sales were in excess of $40 million during 1968.

Programs of planned change in organiza- tion style have received increasing attention from both researchers (Friedlander, 1967) and practitioners (Marrow, 1966 and 1967), but serious gaps exist in knowledge about appropriate learning designs as well as about their consequences (Schein and Bennis, 1965). This study explores both aspects of the problem. The learning design derived from the laboratory approach and used Likert's (1967:196-211) profile of organiza- tional characteristics as the standard for measuring progress toward the planned organization style.

LABORATORY APPROACH TO ORGANIZATION DEVELOPMENT

Basically, the laboratory approach (Brad- ford et al., 1964) to organization develop-

ment seeks to create a specific kind of social order, and provides experience with the skills and attitudes appropriate to that order. The typical learning design has three emphases, whose basic thrust is to use some temporary environment to generate learning which can be applied at the worksite. First, members of an organization learn, by experience in a con- trived social laboratory, that certain norms or values can guide interpersonal and inter- group relations while helping meet important personal and organizational needs. Second, members gain experience during the labora- tory with attitudes and behavioral skills necessary to approach these norms. Third, a substantial cadre of organization members becomes convinced that laboratory norms should guide behavior in organizations, and also that the members can provide or develop attitudes and behavioral skills appropriate to those norms.

The norms or values underlying typical ef- forts toward organizational development are complex, and can only be sketched here. They include (Slater and Bennis, 1964) agreement to work toward (1) full and free communication within and between hier- archical levels; (2) reliance on open con- sensus in managing conflict, as opposed to coercion or compromise; (3) influence based

79

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 3: Planned Change in Organization Style Based on the Laboratory Approach

80 ADMINISTRATIVE SCIENCE QUARTERLY

on competence rather than on personal whim or formal power; (4) respect of norms that permit expression of emotional as well as task-oriented behavior; and (5) acceptance of conflict as a phenomenon that has to be coped with willingly and openly.

Initial experience with these norms and with their associated attitudes and behavioral skills may be gained in a variety of temporary learning environments. The sensitivity train- ing group-or T-group, a small number of people with the goal of learning about one another from one another within a frame- work of values for guiding interpersonal and intergroup relations (Argyris, 1964:60-61) is a particularly useful vehicle for inducing such learning. Experience with appropriate norms and skills can also be gained in other learning contexts, however, such as man- agerial grid training (Blake and Mouton, 1964) or the confrontation design (Golem- biewski and Blumberg, 1967).

The various temporary learning environ- ments derived from the laboratory approach commonly attempt to provide a model for interpersonal and intergroup relations-first within small learning groups and later in large organizations-that emphasizes in- creasing trust and decreasing risk among the participants. The specific goal is to permit the individual to relax his attachment to his old values, attitudes, and behavioral skills, at least enough to experiment with and eval- uate new ones (Bennis, 1966:1-18). Such relaxation is possible and usually exhilarating for most people, although a few people find it unpleasant and fewer still react more negatively. As Mill (1962) has argued, re- gression is an important part of the dynamics of such relaxation in T-groups. If the regres- sion is severe enough, ego disintegration can occur. In this sense, the present design is a very safe one, for the focus is narrowly on managerially relevant dynamics, as opposed to a total life style; participants are not strangers on a cultural island; totally new relations need not be developed among par- ticipants; and so on.

The crucial issue in organizational develop- ment programs is the transfer into large organizations of the initial experience with laboratory norms, attitudes, and behavioral skills that typically has been gained in small groups. The practical difficulties of transfer

are great, but compelling reasons support the attempt. Figure 1 provides an overview of how the laboratory approach to organiza- tional development can provide useful re- turns for both the individual and his organization (Golembiewski, 1969). That overview is based on the guiding belief that, in the long run, the satisfaction and produc- tivity of members will be greatest when in- dividual and organization needs are both met.

Figure 1 requires two caveats. The extent to which individual and organization needs can be met simultaneously will, of course, always be a problem. The laboratory ap- proach to organizational development at least attempts to institutionalize processes, under- lying attitudes, and behavioral skills that facilitate the goal in organizations with com- plex and changing missions and roles. Also, there is no unanimity about the optimistic rationale for the use of the laboratory ap- proach. The present purpose is not to defend that rationale, however. Rather the objec- tives here are to describe a learning design based on that rationale, to test that design, and thus begin to assess the credibility of the rationale.

GOALS OF PLANNED CHANGE The program reviewed here had 3 short-

run and 3 long-run goals. They were a com- bination of operating needs, as perceived by the authors and by the managers who com- prised the experimental population, and of consultant strategies intended to meet these needs. Both authors had long-standing ex- perience with the parent organization from which the managers came, and mutual trust was at a high level from the first phases of the program. Acceptance of the authors as external change-agents was both easy and rapid.

Short-Run Goals

First, the paramount short-run goal was to integrate a new management team at the top three levels of a national field-sales organiza- tion (Figure 2). Within the year preceding the program, there had been personnel changes at all three management levels, in- cluding a new national sales manager, a divi- sional sales manager, and three regional sales managers. The goal was to develop effective

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 4: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 81

Basic Premise: When individuals can meet their own needs while meeting organizational needs, member satisfaction will be greatest, and output wvill be qualitatively and quantitatively best.

.

An individual's basic needs center around self- An efficient organization will develop an appro- realization and self-actualization. The former in- priately-shifting balance between institutionaliza- volves a person seeing himself as he is in interac- tion and risk-taking. The former refers to infusing tion with others, with the goal of increasing the with values the activities of the organization, so congruence between his intentions and his impact as to elicit member support, identification, and on others. Self-actualization refers to the processes collaboration. Risk-taking is necessary in innovat- of growth by which an individual realizes his ing more effective ways to deal with existing ac- potential. tivities, and in adapting to environmental changes

in society, markets, technology, and so on.

An individual whose basic needs are satisfied does An organization's successful balancing of institu- not seek comfort and security; rather, he searches tionalization and risk-taking wvill depend upon for work, challenge, and responsibility. (a) the increasingly complete use of people as

J well as non-human resources, (b) the development & W and maintenance of a viable balance between

central control and local initiative, (c) fluid lines of communication, vertically, horizontally, and diagonally, and (d) decision-making processes that solve problems that stay solved without creating other problems.

Satisfaction of both individual and organization needs wvill be facilitated by, if such satisfaction does not in fact crucially depend upon, skill and competence in interpersonal and intergroup situations.

An individual's growth and self-realization are Organizational family teans are exposed to sensi- facilitated by interpersonal relations that are hon- tivity training, with the intention of increasing trust est, caring, and non-manipulative. Hence the re- and responsibility that can be applied directly to liance on stranger experiences in sensitivity training solving organizational issues, and with the inten- groups composed of individuals with no past rela- tion of decreasing the risk in being open in inter- tionships. Such training is a managed process of personal and group situations. Skill and compe- gaining experience with attitudes and skills for tence in interpersonal and intergroup situations inducing greater openness about positive and can be increased in sensitivity training groups com- negative feelings, attitudes, or beliefs. Stuch open- posed of strangers, that is, but the real test is the ness leads to greater trust and reduced risk in com- application of such learning in life-relevant situa- municating in the stranger group, and is intended tions. Such application will require that substantial to suggest possible transfers into other environ- numbers of organization members learn appro- ments. priate interpersonal skills, as well as that they

internalize a set of values which support and x6 I reinforce such learning.

Persons in groups which develop greater openness tend to identify strongly with other members and with the goals of the group

Groups characterized by strong identification with members and goals become increasingly capable of dealing with issues relevant to their members, and hence increasingly capable of influencing their environment in desired ways.

Groups whose members identify strongly and who can influence their environ- ment are likely to be effective reinforcers of decisions about change. Such groups also can provide emotional support necessary to sustain required changes in the values, attitudes, or behaviors of their members.

FIGURE 1. A MODEL OF THE LABORATORY APPROACH

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 5: Planned Change in Organization Style Based on the Laboratory Approach

82 ADMINISTRATIVE SCIENCE QUARTERLY

National Sales Head

(i | Divisional | Divisional l I Manager, Manager,

l l East l West

7 Regional 6 Regional Managers Managers

1 45 1 35 1 l Salesmen Salesmen

FIGURE 2. INTERRELATIONSHIPS WITHIN EX-

PERIMENTAL UNIT

and adaptive top-to-bottom and bottom-to- top relations among all three managerial levels of the experimental unit.

Second, horizontal linkages were to be de- veloped between divisional managers and es- pecially between the regional managers. The regional managers had been spending about 60 percent of their time in direct selling, but would rapidly transition to perhaps 90 per- cent of their time in managerial work, coin- cident with the introduction of a radically different product line. Changes in both roles and product required complex and unparal- leled cooperative and innovative effort. The underlying concept of the planned change was the creation of three interlocking teams, as in Likert's (1961:107-115) notion of the linking pin. This concept required building a sense of team identification within hierar- chical levels, as well as between them, and implied rejection of the common strategy of building cohesiveness in each of the divi- sional units by inducing win-lose competition between them.

Third, the planned change was intended to help confront and resolve some nagging

problems of personal style and organization history. For example, the new national sales head was widely seen as so aggressive and punishing that his subordinates were afraid of him. The regional managers feared that their interests would not be accurately repre- sented to him, their expectation being that the head would dominate the divisional man- agers. The divisional managers themselves had similar concerns. Consequently, upward communication was inhibited at the time when it was most vital. Furthermore, com- mon opinion held that the regional managers had been poorly and even harshly managed over an extended period. The sales head and the two divisional managers individually were anxious to make changes, but they were concerned that almost any change would be interpreted by their subordinates as callous or vengeful, or worse. Finally, several of the regional managers had been involved in the organizational scuffling for the two top jobs, either as competitors or as supporters of com- petitors. Consequently, relations between peers were stressful when goodwill and trust were greatly needed.

Long-Run Goals

The first long-run goal of the program was to increase the congruence between the be- haviors that the prevailing organization style demanded of the managers and the behaviors that the men preferred. For example, all of the managers feared that the existing re- lations in the experimental unit called for substantial mistrust and secrecy, while the managers themselves preferred to be more trusting and open.

A second long-run goal was a change in organizational style to permit greater congru- ence of individual needs and organization demands. Roughly, the goal was movement toward an open, adaptive, and problem- solving system such as Shepard (1965) de- scribed, in contrast to a closed, coercive, and hierarchy-serving system. The experimental unit initially was perceived by the consul- tants as somewhere between these polar ex- tremes, but under the new managers, the unit was widely seen by their subordinates as threatening to move significantly in the closed direction.

The third goal of the program was to ex-

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 6: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 83

periment with a bottom-up approach to organization development, instead of the more typical top-down approach. Change agents usually direct their interventions toward the top of the organizational hier- archy, or as near to it as possible, for good reasons. However, this top-down strategy often produces resistance at lower levels, and it certainly limits points of access by external change-agents. In any case, the present goal was to gain knowledge and experience about organizational conditions and learning de- signs that facilitate change at the middle levels of a large organization.

The basic learning design attempted to meet two imperatives: managers were en- couraged to consider a change in the style of their interpersonal and intergroup rela- tions, but they were continually alerted to the constraints on what they and their im- mediate superiors could reasonably influence. Consequently, the learning design featured a number of decision points explicitly intended to test commitment by various levels of man- agement to the unfolding program. The con- sequence to be avoided was a feeling by any of the managers that they had been manipulated into a position they found un- comfortable, but from which they could not gracefully escape. In positive terms, the over- all objective was a feeling of psychological success among the managers: a sense of personally owning the change program as opposed to a sense of the program being im- posed upon them.

SENSITIVITY TRAINING IN ORGANIZATION DEVELOPMENT

The learning design used the laboratory approach (Bradford et al., 1964) in a variety of ways. First, the design emphasized the use of T-groups. T-groups provided the initial experience with, and developed skills and at- titudes appropriate for, values to guide inter- personal and intergroup relations that differ from the pyramidal values commonly ac- cepted by members of large organizations (Argyris, 1964:60-61). Second, the design attempted to suggest how laboratory values and dynamics could be built into more com- plex organizations, as by improving feedback processes in large units. Third, the design at- tempted to motivate organization members to

evolve a set of organizational values that would reinforce and encourage building analogs of laboratory dynamics into their work units.

These general motivations derive specific content from four classes of values that char- acterize the laboratory approach, as its appli- cations cover the range from T-groups to complex organizations. The four classes of laboratory values and their characteristics may be outlined briefly:

Meta-values of laboratory training (Schein and Bennis, 1965; Bradford et al., 1964) 1. An attitude of inquiry reflecting (among

others) a willingness to analyze and to experiment with behavior.

2. An expanded consciousness about the effects of behavior and a sense of in- creased alternatives available to the in- dividual.

3. The value system of democracy, stressing a spirit of collaboration and open resolu- tion of conflict through a problem-solving orientation.

4. Emphasis on mutual helping relationships as the best way to express the interde- pendency of people.

Proximate goals of laboratory training

1. Increased insight and self-knowledge. 2. Sharpened diagnostic skills at all levels,

that is, on the levels of the individual, group, organization, and society.

3. Awareness of, and practice of skills in creating, conditions of effective function- ing at all levels.

4. Testing self-concepts and skills in inter- personal situations.

5. Increased capacity to be open, to accept feelings of self and others, and to risk interpersonally in rewarding ways.

Desirable means for laboratory training

l. Emphasis on here-and-now occurrences. 2. Emphasis on the individual act rather

than on the total personal acting. 3. Emphasis on feedback that is noneval-

uative in that it reports the impact on the self of other's behavior, rather than feed- back that is judgmental or interpretive.

4. Emphasis on unfreezing behaviors the

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 7: Planned Change in Organization Style Based on the Laboratory Approach

84 ADMINISTRATIVE SCIENCE QUARTERLY

trainee feels are undesirable, on practice of replacement behaviors, and on rein- forcement of new behaviors.

5. Emphasis on trust in leveling, on psycho- logical safety on the trainee.

6. Emphasis on creating and maintaining an organic community of mutual helpers.

Organization values consistent with'i labora- tory training

They are the 5 goals discussed in the section above: Laboratory approach to organization development.

Generally, these values were stressed in the learning design, and appropriate behaviors were practiced. For example, exercises with giving and receiving feedback were designed to highlight the prescriptions listed above as desirable means for laboratory training, and these exercises encouraged the development of attitudes and behavioral skills appropriate to those values. The flow of the learning design (Golembiewski and Blumberg, 1969) successively emphasized the four classes of values.

The present use of the laboratory approach had five other features. First, the final de- cision to adopt the approach was reached through planned stages. The national sales head and divisional managers individually had sensitivity training experiences to give them some basis for deciding whether or not to authorize a program of organization devel- opment based on the laboratory approach. Al three managers were attracted to the ap- proach, based on their own independent re- actions and those of 65 other members of their parent organization who had had T- group experiences in other company pro- grams. In addition, the regional managers were given an opportunity to express their reactions to a program based on the labora- tory approach, in the form of a discussion of a descriptive article about T-groups (Argyris, 1964). Any strong resistance by the regional managers would at least have delayed the program. Both the three top managers and the consultants were careful not to overload the already tense relations between man- agerial levels. The regional managers agreed to participate, and the program began.

Second, the initial phase of the learning

design was a T-group experience for the re- gional managers, at which the consultants served as trainers. This initial phase lasted four-and-a-half days, at a resort site. The overall goal was to improve interpersonal and intergroup relations in the organization, not necessarily to enhance personal growth or to provide an emotionally-moving experience, and this goal structured the experience in significant ways. For example, the focus limited both the topics and the emotional depth of discussion, as compared with T- groups composed of strangers. The regional managers were told that the T-group was preparation for the day-and-a-half following it, when they could confront the two divi- sional managers with their concerns about interpersonal and intergroup relations in the experimental unit. The words could confront were emphasized. The consultants had only arranged with the two division managers that they be available on short notice, should their presence seem appropriate, and this arrange- ment was shared with the regional managers as a first piece of business in their T-group. The regional managers quickly and correctly perceived it as an escape hatch should a con- frontation with superiors seem too risky to them or inappropriate to the consultants.

Building toward the possible confrontation with superiors as a climax, the regional managers were encouraged to explore their feelings and reactions, both about the orga- nization and the T-group experience. The dis- cussion shifted between emphasis on here- and-now socio-emotional dynamics triggered in the T-group and on there-and-then orga- nizational issues. The consultants used that discussion to help the regional managers be- come more aware of their feelings and reac- tions, and to attach these reactions more precisely to the stimuli that induced them. Such sensitivity was seen as crucial prepara- tion for the possible confrontation with the divisional managers, when a premium would be put on how and why the regional man- agers were reacting and feeling. The regional managers tested how rewarding for them openness was in their T-group, and con- sciously weighed the openness they were willing to risk in the confrontation with the two divisional managers. In addition, the T- group provided experience with attitudes and

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 8: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 85

skills that would facilitate the confrontation between superior and subordinate. The dis- cussion permitted a cross validation of con- cerns and reactions about organization rela- tions, so that each regional manager would know as clearly as possible what his colleagues were thinking and feeling. Again the purpose was to facilitate the subsequent confronta- tion. For example, consensual validation among the managers about what the real issues were would reduce the risk to individ- ual regional managers of confronting their superiors later.

Third, the regional managers proved eager for the confrontation. During the confronta- tion phase, they met either as two divisions, each with their superior, or as a joint group with both divisional managers present. Only minimal structuring cues were provided for the regional managers. They were given a half-day to decide whether to hold the con- frontation and to plan it, following the model developed in Golembiewski and Blumberg (1967). The confrontation was described by the consultants as an opportunity for the re- gional managers to test the degree to which the national sales head and the two divisional managers really wanted an open, problem- solving system. The risks to the division man- agers received careful attention. From the start, the design stressed the risks implied in the confrontation; our motto was drawn from the experience of the Crusaders: "More went than came back." Instructions also under- scored the need to concentrate on issues that were within the competence of superiors, or could be influenced by them.

The two divisional managers presented a more difficult problem. Their anxiety levels were high; they were very conscious of being outnumbered; and they did not know what they might expect from their subordinates. The consultants attempted to work through such concerns with them in a briefing session immediately before the confrontation. The briefing was facilitated by the fact that the consultants had been the trainers in the T- groups previously attended by the two divi- sional managers. The divisional managers were instructed only to try to understand any feedback that the regional managers offered, to explore both ideas and feelings with as little defensiveness as possible, and to try to

do so without projecting an attitude of "ca- tharsis is good for them." The consultants were available during the confrontation, to provide support and to reinforce adherence to the norms of the laboratory approach.

Fourth, following the confrontation, the di- visional managers met individually with the national sales head to clarify their own rela- tions, in part in response to the confrontation. This phase was preplanned and announced to the regional managers as a possibility in the introductory session of the training design.

Fifth, process observers, who were trained company employees, attended subsequent meetings at worksites of managers at all three levels. Their position as observers symbolized the transfer of the guardianship of the lab- oratory values from outside consultants to agents of the organization. Process observa- tion was intended to reinforce laboratory values in action settings, both by the presence of the observer and through his interventions, which encouraged looking at the socio-emo- tional processes of the meeting as well as its content (Bennis and Shepard, 1956). The goal was to preclude the initial off-site train- ing experience being seen as a magic moun- tain phenomenon, unique and perhaps precious, but remote from the workaday world.

In sum, the learning design sought both to initiate attitudinal and behavioral change in a temporary learning environment and to transfer that change into the organization. The T-group experiences were meant to re- lax old attitudes about interpersonal and intergroup relations. The confrontation pro- vided an opportunity to practice new atti- tudes and behaviors under actual conditions, as well as to test the risk involved for all. And process observers at subsequent work meetings of the regional managers reinforced the new behavioral skills and attitudes in organizational settings.

MEASURING PROGRESS TOWARD PLANNED CHANGE

Progress toward planned change in the or- ganization style of the sales unit was mea- sured by self-reported changes in the 48 items of Likert's profile of organizational character- istics. Six properties of the profile require emphasis. First, each item is represented by

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 9: Planned Change in Organization Style Based on the Laboratory Approach

86 ADMINISTRATIVE SCIENCE QUARTERLY

System of Organization

System 1 System 2 System 3 System 4

Exploitative- Benevolent- Consultative Participative Authoritative Authoritative Group

Bulk of decisions at Policy at top, many Broad policy deci- Decision making top of organization decisions within sions at top, more done throughout

prescribed frame- specific decisions organization, al- work made at at lower levels though well in- lower levels but tegrated through usually checked linking process with top before provided by over- action lapping groups

Item 33: At what level in the organization are decisions formally made?

FIGURE 3. REPRESENTATIVE ITEM FROM LIKERT'S PROFILE

a 20-point scale anchored by four descrip- tions, as is illustrated in Figure 3.

Second, Likert distinguishes the four major intervals as different systems of organization, with the participative-group system being most consistent with the laboratory approach. Therefore, scores on the 48 profile items can provide estimates of the degrees to which the style of an organization approaches the norms of the laboratory approach (Marrow et al., 1967:45-60).

Third, responses were uniformly coded so that scores run from 1 for the extreme ex- ploitative-authoritative rating to 20 from the extreme participative-group rating. The Li- kert scales vary in direction, so as to inhibit a response set; however, all discussion below is based on the uniform coding of responses irrespective of the scale directions of items.

Fourth, each profile item was rated twice: once as ideal, as respondents felt their orga- nization unit should be; and also as now, as respondents actually saw their organization unit to be. All responses were sent directly to the senior author at his university address, with guarantees that only aggregate data would be reported. Responses from individ- ual managers were identified by name, which made possible matched comparisons of the responses of each manager on test 1 and test 2.

Fifth, the profile taps a range of processes that determine an organization's style. They are: (1) leadership, (2) character of moti- vational forces, (3) character of communica- tion, (4) character of interaction-influence, (5) character of decision-making, (6) char- acter of goal setting or ordering, and (7) character of control.

Sixth, respondents were instructed to think of the experimental unit as a whole in mak- ing their ratings. The data treated in this paper are the self-reports by the regional managers.

The profile can serve both descriptive and predictive uses. Descriptively, responses to the profile items could have been used to estimate the initial climate of the experimen- tal unit. Ideal scores before the learning ex- perience were expected to cluster in the System 4 area of the profile, and most now scores were expected to fall between Systems 2 and 3. Administration of the profile before the learning experience could have helped judge the cultural preparedness of the sys- tem for the laboratory approach. For ex- ample, a System 1 organization is probably not suitable for an organizational develop- ment program based on the laboratory ap- proach. In the present case, the consultants relied on their intuition and knowledge of the parent organization when making the

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 10: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 87

moral judgment to recommend the laboratory approach.

Likert's profile also can be used predic- tively. The goals of the organizational devel- opment effort were defined in terms of an- ticipated changes in scores on the profile between two administrations of the question- naire, given a successful learning design. Test 1 came early in the week of training; and test 2 followed four months later. If the learning design had the intended effects, comparisons of the two tests should support the following expectations:

1. The original state of interpersonal and intergroup relations in the sales unit would fall significantly short of the expectations of participants; i.e., Ideal 1 scores would be sig- nificantly greater than Now 1 scores.

2. The intervention would raise partici- pants' expectations about how need-gratify- ing their organization could be expected to be; i.e., Ideal 2 scores would be significantly greater than Ideal 1.

3. The intervention would significantly change the organization style in the directions that participants preferred; i.e., Now 2 scores would be significantly greater than Now 1 scores.

4. A successful intervention need not eliminate the differences between the existing interpersonal and intergroup climate and the expectations of organization members; i.e., Ideal 2 scores may still differ significantly from Now 2.

5. Although a successful intervention need not eliminate them, the differences between scores of the existing climate and scores of the preferred climate should at least remain constant, if not decrease, so as to avoid dif- ferences that might become so large as to be demotivating; i.e., at a minimum (Ideal 2- Now 2) differences would be no larger than (Ideal 1-Now 1) differences.

It was not possible-given the nature of the organization and the small managerial pop- ulation-to employ a control group in the research design. Two points provide detail about this major limitation. The entire man- agerial population was exposed to the ex- perimental treatment. This was possible due to the small number of managers in the ex- perimental unit, and it also was convenient

because of features of the learning design that argued against exposing only a part of this small number of managers to the labora- tory (Golembiewski and Blumberg, 1969). Relatedly, the use as a control of a matched group of managers at similar levels from throughout the organization was considered inappropriate. The experimental unit was unique in many senses: it was adding a radically different product line; it had expe- rienced a major managerial succession; and its history had a strong and distinctive pat- tern.

Although the lack of a control group does require a definite tentativeness in attributing any results to the training design, it does not seem likely that the effects reported were due to the passage of time or to the three new managers alone. Basically, the changes in self-reports in the experiment are contrary to the history of the experimental unit as well as to the previous styles of the national sales head and the divisional managers. Indeed, organization members tended to agree that time alone would have induced movements on the Likert dimensions opposite to those observed. Furthermore, the changes in self- reports were marked and consistent with the broad theory underlying the laboratory method. This is only suggestive of the po- tency of the training design, but at a min- imum, it does urge further research dealing with organizational applications of the lab- oratory approach (Schein and Bennis, 1965: 323).

Other significant factors also must temper interpretation of changes in self-reports about organization style, until replicatory studies are made. A kind of Hawthorne effect may explain the results. That is, changes in self- reports might be due not so much to the training design as to the attention given the regional managers by consultants and top management. No one can conclusively reject a Hawthorne effect in this case. The general consistency of the results with the theory underlying the training design argues against a Hawthorne effect determining the results, but only replications of the present training design can test this objection. More seriously, the results below may not reflect actual be- havorial change. Trainees may have learned the kinds of behaviors that the consultants

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 11: Planned Change in Organization Style Based on the Laboratory Approach

88 ADMINISTRATIVE SCIENCE QUARTERLY

valued; they may have dutifully reported ap- propriate changes in their self-reports as co- operative subjects in an experiment often do (Webb et al., 1966:10-32); yet the style of their organization might not have changed at all. Similarly, change in self-reports may be due in part to the practice gained in respond- ing twice to the Likert form (Cane and Hein, 1950) or to reliability problems with the instrument.

The possibility that the changes reported below do not reflect actual behavioral changes must be taken especially seriously, even though a plausible counter-case can be built. Substantial anecdotal evidence sug- gests that extensive behavioral change did take place. In addition, rating each Likert dimension as ideal and now was an attempt to force respondents to distinguish between wish-fulfillment and reality. Various technical details-such as spacing the two administra- tions of the Likert profile four months apart, and varying the direction of scales for in- dividual Likert items to prevent a response set-at least made it more probable that respondents would base their reports on ac- tual behavior. However, the present research does not deal directly with behavioral change, and the possibility of a subtle kind of wish-fulfillment cannot be rejected.

SOME RESULTS OF THE PLANNED INTERVENTION

Despite these reservations about the data, the responses on two administrations of the Likert profile imply that significant changes in the climate of the experimental unit oc- curred in the 4-month interval after the learning experience. In general, these changes in self-reports imply that managers were more free to behave at work in ways they preferred, and that the managers' work cli- mate became more what they wished it to be. Table 1 suggests the gross magnitude of the changes.

The specific changes in self-reports can be established by an interpretive review of the data reported in Table 2. Each of the five columns in Table 2 tests one of the expecta- tions concerning the effects of a successful learning design. Due to space limitations, the profile items in Table 2 are not accompanied by the four descriptive statements that an-

TABLE 1. COMPARISON OF SCORES ON TWO

TESTS OF THE PROFILE

Test 1 Test 2

System Ideal Now Ideal Now

Exploitative- Authoritative 0 0 0 0

Benevolent- Authoritative 0 1 0 0

Consultative 3 41 1 18 Participative

Group 45 6 47 30

chor the 20-point scale for each item. The 1-Ni column shows the initial incon-

gruence of the needs of the managers and the organization style that they reported as ex- isting before the planned intervention. Ideal 1 is greater than Now 1 in all 48 cases, and 43 of these differences reach the .05 level of statistical significance, or better. This pro- vides impressive support for expectation 1. For example, for item 1, 11-N1 indicates that participants desired significantly more trust and confidence than the organization style encouraged.

The data in the 12-11 column are consistent with expectation 2, suggesting that the inter- vention helped significantly change what the participants expected of their organization. The changes from Ideal 1 to Ideal 2 scores were positive in 42 of 48 cases, with 19 of these reaching the .05 level or better. None of the 6 negative changes approach usually accepted levels of significance. This change in ideal scores is important because the heightened expectations can be expected to induce motivational forces to change the or- ganization style so as to reduce the discrep- ancy between ideal and now scores. Such an effect can be expected, except where the gap between ideal and now scores is too great.

The N2-N1 column supports expectation 3 and its imputation of motivational force to increases in ideal scores. That is, comparison of Now 2 scores with Now 1 demonstrates that the participants' descriptions of the ex- perimental unit reflected movement toward an organization style consistent with the lab- oratory approach; N2 is greater than N1 in 47 of the 48 comparisons, with 28 reaching the .05 level of significance, or better. An

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 12: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 89

TABLE 2. MEAN DIFFERENCES BETWEEN COMBINATIONS OF PROFILE SCORES

(2 - N2) - Profile items I1 -N1 '2'-1 N2-N1 12-N2 (11-N1)

Leadership Processes

1. Extent to which superiors have confidence and trust in subordinates 5.25* 0.25 3.00' 2.50* - 2.75*

2. Extent to which subordinates, in turn, have confidence and trust in superiors 5.00* 1.25 3.50* 2.75* - 2.25*

3. Extent to which superiors display suppor- tive behavior toward others 3.58* 3.17* 3.33* 3.42* - 0.17

4. Extent to which superiors behave so that subordinates feel free to discuss important things about their jobs with their immediate superior 4.83* 1.42f 2.42+ 3.83* - 1.00

5. Extent to which immediate superior in solving job problems generally tries to get subordinates' ideas and opinions and make con- structive use of them 5.33* -0.17 2.75f 2.42* - 2.92*

Character of Motivational Forces 6. Underlying motives tapped 5.25* 0.58 2.25+ 3.58* -1.67 7. Manner in which motives are used 2.92* 0.08 1.67+ 1.33+ -1.58 8. Kinds of attitudes developed toward or-

ganization and its goals 3.17* 1.58f 1.92+ 2.83* -0.33 9. Extent to which motivational forces con-

flict with or reinforce one another 5.83* 1.83f 4.42f 3.25* -2.58f 10. Amount of responsibility felt by each

member of organization for achieving organiza- tion's goals 2.67+ 2.00 1.00 3.67* 1.00

11. Attitudes toward other members of the organization 2.92* 1.33 0.50 3.75* 0.83

12. Satisfaction derived 3.67* 0.83 2.25+ 2.25* -1.42

Character of Communication Process 13. Amount of interaction and communica-

tion aimed at achieving organization's objectives 4.67* 1.25* 2.75+ 3.17* - 1.50 14. Direction of information flow 4.75f 0.83 2.25 3.33* -1.42 15. Where downward communication ini-

tiated 3.67f 2.58f 3.83* 2.42* -1.25 16. Extent to which superiors willingly share

information with subordinates 3.08* 1.92f 3.25f 1.75f -1.33 17. Extent to which communications are

accepted by subordinates 1.67 2.42+ 1.17 2.92* 1.25 18. Adequacy of upward communication via

line organization 4.67* 1.58* 2.83+ 3.42* -1.25 19. Subordinates' feeling of responsibility for

initiating accurate upward communication 3.67* 1.17f 0.67 4.17* 0.50 20. Forces leading to accurate or distorted

upward information 2.33* 1.08 0.58 2.83* 0.50 21. Accuracy of upward communication via

line organization 4.00f 0.42 1.67t 2.75* -1.25 22. Need for supplementary upward com-

munication system 1.17 2.25 2.67+ 0.75 -0.42 23. Sideward communication, its adequacy

and accuracy 5.58* 0.75 3.33-f 3.00* -2.58 24. Psychological closeness of superiors to

subordinates (i.e., friendliness between superiors and subordinates) 1.58 4.17* 2.42+ 3.33* 1.75

+ Difference statistically significant at .05 level by K-test. f Difference statistically significant at .01 level by K-test. * Difference statistically significant at .001 level by K-test.

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 13: Planned Change in Organization Style Based on the Laboratory Approach

90 ADMINISTRATIVE SCIENCE QUARTERLY

TABLE 2 (Continued)

(12 - N2) -

Profile items 11-N1 I2-It N2-N1 12-N2 (11-N1)

25. How well does superior know and under- stand problems faced by subordinates? 3.83f 1.00 1.42 3.42* -0.42

26. How accurate are the perceptions by superiors and subordinates of each other? 5.42* 0.83 3.00* 3.25* -2.17f

Character of Interaction-Influence Process 27. Amount and character of interaction 4.42* 1.50 1.83t 4.08* -0.33 28. Amount of cooperative teamwork present 2.18 1.45t 1.00 2.75* 0.45 29. Extent to which subordinates can influ-

ence the goals, methods, and activity of their units and departments, as seen by superiors 4.25* 1.42t 3.17f 2.50* - 1.75f

30. Extent to which subordinates can influ- ence the goals, methods, and activity of their units and departments, as seen by subordinates 4.67* 0.75 1.50 3.92* -0.75

31. Amount of actual influence which su- periors can exercise over the goals, activity, and methods of their units and departments 4.33* -0.58 2.00 1.75+ -2.58+

32. Extent to which an effective structure exists enabling one part of organization to exert influence upon other parts 4.83* 2.00t 1.75 5.08* 0.25

Character of Decision-Making Process 33. At what level in organization are deci-

sions formally made? 5.00f 1.27 3.67t 3.09* -2.36 34. How adequate and accurate is the in-

formation available for decision making at the place where decisions are made? 4.92* -1.33 0.50 3.08* - 1.83t

35. To what extent are decision makers aware of problems, particularly those at lower levels in the organization? 4.17f 0.17 1.67 2.67f -1.50

36. Extent to which technical and profes- sional knowledge is used in decision making 5.33* 1.33 3.254 3.42* -1.92

37. Are decisions made at the best level in the organization as far as availability of the most adequate information bearing on the decision? 2.25t -1.58 -1.58 2.25+ 0.0

38. Does the decision-making process help to create the necessary motivations in those persons who have to carry out the decisions? 4.83* 0.08 3.25f 1.67t -3.17f

39. To what extent are subordinates in- volved in decisions related to their work? 5.91* 0.20 2.50t 3.73* -2.70

40. Is decision making based on man-to- man or group pattern of operation? Does it encourage teamwork? 2.36+ 1.20+ 0.10 3.55* 1.10

Character of Goal Setting or Ordering 41. Manner in which goal setting or order-

ing is usually done 3.82f 2.40t 2.90t 3.09* -0.50 42. To what extent do the different hierar-

chical levels tend to strive for high performance goals? 1.27 1.50 1.50 1.45+ 0.0

43. Are there forces to accept, resist, or reject goals? 3.36* 2.50+ 2.70+ 3.18* -0.20

Character of Control Process 44. At what hierarchical levels in organiza-

tion is there major or primary concern with the performance of the control function? 4.36f -0.20 1.00 3.45* - 1.20

45. How accurate are the measurements and information used to guide and perform the con-

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 14: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 91

TABLE 2 (Continued)

(I2 - N2) - Profile items I1-N1 I2'I1 N2-N1 I2-N2 (11-No) trol function, and to what extent do forces exist in the organization to distort and falsify this information? 2.64+ 0.70 0.40 2.27* 0.30

46. Extent to which the review and control functions are concentrated 3.45f 1.10 1.20 3.09' -0.10

47. Extent to which there is an informal organization present and supporting or opposing goals of formal organization 3.91' 0.60 1.70 2.55* -1.10

48. Extent to which control data (e.g., ac- counting, productivity, cost, etc.) are used for self-guidance or group problem solving by man- agers and non-supervisory employees, or used by superiors in a punitive, policing manner 4.64* 0.90 2.90t 2.36* 2.00

additional 10 cases approach the .05 level. With both ideal scores and now scores in-

creasing in the interval between tests, the question arises as to whether changes in or- ganization style ran fast enough to stand still, as it were, relative to the increasing expecta- tions implied by increasing ideal scores. The danger is that the learning design merely increased the gap between ideal and now scores. Up to a point, such a gap has the motivational properties associated with a de- manding but still attainable level of aspira- tion. However, the gap could grow so great as to create doubts that the level of aspiration can ever be attained. Such gaps would prob- ably curb effort to attain the goals, and would have demotivating properties.

Expectations 4 and 5 attempt to take such issues into account. They are meant as boundary conditions for defining a successful learning design. Expectation 4 holds that a successful intervention is not defined in terms of eliminating I-N gaps; indeed, that might sacrifice motivational forces. This is a per- missive criterion of success in the learning design. However, expectation 5 provides that the I-N gaps should at least not be in- creased substantially by the learning design. The point of this boundary condition is a conservative one: to avoid increasing I-N gaps beyond that unknown point at which they become demotivating. This is a strict criterion of the success of the learning design.

The 12-N2 column in Table 2 establishes that the learning design did not eliminate I-N differences, which meets our permissive expectation 4. Considering only the second

test, 12 is greater than N2 in all 48 cases, and 47 of these cases reach at least the .05 level of statistical significance. This pattern is similar to that generated by comparisons of ideal scores and now scores for test 1.

There is no easy interpretation of what these gaps in ideal versus now scores mean; significant differences between ideal and now scores on test 2 do not mean that nothing hap- pened as a result of the training design. How to evaluate what the data do mean is more of a problem, however. Certainly, the data in the 12-N2 column mean that the organiza- tion is dynamically different from what it was. Certainly also, the patterns of ideal and now scores imply one danger of organizational development programs. The learning design apparently created rising expectations, and these increased expectations will decrease the alternatives available to management. Should future changes in organization style move toward systems that are closed, coercive, and hierarchy-serving, the increased ideal scores imply a heightened resistance by managers. An organization development program based on the laboratory approach, in this sense, seems to set a tiger loose in the streets.

The last column suggests that the gaps be- tween ideal and now scores have been re- duced, even though those gaps are still significant. This meets expectation 5 for a successful change. Since ideal is almost always greater than now, (12-N2) - (11-N1) should be negative if there were a decrease in the gap between ideal and now scores. This is indeed the case: 36 of the 48 comparisons in the last column are negative. Although only

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 15: Planned Change in Organization Style Based on the Laboratory Approach

92 ADMINISTRATIVE SCIENCE QUARTERLY

8 of these negative values reach the .05 level, an additional 15 cases closely approach that level, and 2 cases show no difference. Less than one case in five reflects an increased gap between ideal and now scores, and not one of these positive changes approaches usually accepted levels of significance.

The organizational development effort did not merely raise aspirations about what the organization climate should be, thereby mak- ing the gap between ideal and now scores so great as to produce pessimism. Ideal scores did increase, but basic changes in or- ganization style were also encouraged by the learning design.

CONCLUSION

The learning design to change the organi- zation style seems, in general, to have had the intended effects. The design appeared to induce or reinforce appropriate norms in the experimental unit, and these norms were reported to have changed attitudes as well as behaviors over the four-month period of observation.

The results also suggest some significant issues, to which no definitive response can yet be given. Thus interpretation of the present results is somewhat chancy because of the lack of a control group and because of the reliance on self-reports. The total pop- ulation of managers was subjected to the ex- perimental condition and, although it seems very improbable, the changes indicated in Table 2 may reflect environmental changes, or the mere passage of time, more than the products of the learning design. The use of self-reports raises greater questions, with which subsequent research must cope. For example, data are being gathered from the salesmen to cross-validate the changes im- plied in the self-reports of the managers.

Moreover, the persistence of the changes reflected in Table 2 must be established. A third administration of the Likert profile is planned to test for such persistence nine months after the original learning experience. These data will be of great significance in judging the success of this induction of change.

Finally, even granting that the organiza- tional style of the experimental unit was changed in ways that will prove more satisfy-

ing to its members, have the effectiveness and efficiency of the experimental unit been increased? There are no data to answer this question, largely because the missions and roles of the experimental unit have changed rapidly. Participants report more involvement and satisfaction at work; and they generally agree that the style changes will enhance their total performance in the long-run. In addition, sound theoretical reasons urge that high output will co-exist with high satisfac- tion of interpersonal and intergroup needs. But such considerations stop far short of demonstrating that the present change had favorable consequences on output variables (Barnes and Greiner, 1964).

Robert T. Goi'embiewski is research pro- fessor of political science and management at the University of Georgia, and acting head of the political science department. Stokes Car- rigan is head of the training section at Smith, Kline & French Laboratories.

REFERENCES

Argyris, Chris 1964 "T-group for organizational effective-

ness." Harvard Business Review, 42: 60-74.

Barnes, L. B., and Larry E. Greiner 1964 "Breakthrough in organization devel-

opment: Part II." Harvard Business Review, 22:139-165.

Bennis, Warren G. 1966 Changing Organizations. New York:

McGraw-Hill. Bennis, Warren G., and Herbert A. Shepard

1956 "A theory of group development." Hu- man Relations, 4:415-437.

Blake, Robert T., and Jane S. Mouton 1964 The Managerial Grid. Houston: Gulf.

Bowers, David G., Alfred D. Marrow, and Stan- ley E. Seashore

1967 Management by Participation. New York: Harper and Row.

Bradford, Leland P., Jack R. Gibb, and Ken- neth D. Benne

1964 T-Group Theory and Laboratory Method. New York: Wiley.

Cane, V. R., and A. W. Heim 1950 "The effects of repeated testing: III."

Quarterly Journal of Experimental Psychology, 2:182-195.

Golembiewski, Robert T. 1969 "Organization development in public

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions

Page 16: Planned Change in Organization Style Based on the Laboratory Approach

Golembiewski and Carrigan: ORGANIZATION STYLE 93

agencies." Public Administration Re- view, 29:367-378.

Golembiewski, Robert T., and Arthur Blumberg 1967 "Confrontation as a training design in

complex organizations." Journal of Ap- plied Behavioral Science, 3:525-547.

1969 "Sensitivity training in cousin groups: a confrontation design." Training and Development Journal, 23:18-23.

Likert, Rensis 1961 New Patterns of Management. New

York: McGraw-Hill. 1967 The Human Organization. New York:

McGraw-Hill. Marrow, Alfred D.

1966 "Managerial revolution in the state department." Personnel, 43:1-13.

Mill, Cyril R. 1962 "A theory for the group dynamics

laboratory milieu." Adult Leadership, 11:133-134 and 159-160.

Schein, Edgar H., and Warren G. Bennis 1965 Personal and Organizational Change

Through Group Methods. New York: Wiley.

Shepard, Herbert A. 1965 "Changing interpersonal and inter-

group relations in organizations." In James G. March (ed.), Handbook of Organizations: 1122-1135. Chicago: Rand McNally.

Slater, Philip E., and Warren G. Bennis 1964 "Democracy is inevitable." Harvard

Business Review, 42:51-59.

This content downloaded from 188.72.126.118 on Thu, 12 Jun 2014 12:54:22 PMAll use subject to JSTOR Terms and Conditions