Homework answers / question archive / Describe the decision that management had to make, and also whether their decisionmaking style was autocratic, consultative, or group-based

Describe the decision that management had to make, and also whether their decisionmaking style was autocratic, consultative, or group-based

Business

Share With

Describe the decision that management had to make, and also whether their decisionmaking style was autocratic, consultative, or group-based. Explain your reasoning. 2. Go through Vroom-Yetton’s seven questions and apply these questions to their diagram to see what decision-making style their model suggests. Briefly explain the reasoning behind your answers to the seven questions. 3. Conclude your paper with a discussion of whether or not you would recommend this model to your supervisor and whether or not you were surprised by the recommendation given by this model regarding decision-making style. Vroom's contingency model shows that the appropriate leadership style varies with the situation. Moreover, he demonstrates how the leader can learn (1) how to recognize the requirements of a situation and (2) how to tailor his style to meet these requirements. Uan l^eaoers J^earn io J^ecwv Victor H. Vroom ike my fellow authors, I start with certain preconceptions. These preconceptions—some may call them biases—influence the way in which I view issues of leadership, particularly leadership training. I have tried to depict these preconceptions in Figure 1. The central variable in this figure is the behavior of the leader, which I believe is determined by two classes of variables, attributes of the leader himself and attributes of the situation he encounters. Furthermore, I assume that many of the differences in the behavior of leaders can be explained only by examining their joint effects, including interactions between these two classes of variables. The left-hand portion of the diagram is the descriptive .side of the leader behavior equation. Much of my research has focused on these relationships in an attempt to understand the ways in which managers actually respond to situations that vary in a number of dimensions. If you examine the right-hand side of Figure 1, however, you encounter issues that are potentially normative or prescriptive in character. They deal with the consequences of leader behavior for the organi- zation and here I share with Fiedler (and probably disagree with Argyris) a conviction that a contingency model is required. I do not see any form of leader behavior as optimal for all situations. The contribution of a leader's actions to the effectiveness of his organization cannot be determined without considering the nature of the situation in which that behavior is displayed. WORKING WITH THE CONTINGENCY MODEL I am going to assume that most of you arc familiar with the model that Phil Yetton and I developed and have described in detail in our recent book. As a normative model, it deals with the right-hand side of Figure 1, but it is a limited model because it deals with only one facet of leadership behavior—the extent to which the leader shares his decisionmaking power with his subordinates. Figure 2 shows the latest version of our model. For purposes of simplicity, the presentation here is restricted to the model for group problems, that is, problems or deci- This article is based on a paper delivered at the 1975 annual meeting oj the American Psychological Association. Figure 7 SCHEMATIC REPRESENTATION OF VARIABLES USED IN LEADERSHIP RESEARCH 1 Organizational Effectiveness Situational Variables. Leader Behavior 1a Personal Attributes 18 sions that affect all or a substantial portion of the manager's subordinates. At the top of the figure are problem attributes—that is, situational variables that ought to influence the decision process used by the leader—specifically, the amount of opportunity that the leader gives his subordinates to participate in the making o£ a decision. To use the model, one first selects an organization problem to be solved or decision to be made. Starting at the left-hand side of the diagram, one asks oneself the question pertaining to each attribute that is encountered, follows the path developed, and finally determines the problem type (numbered 1 through 12). This problem type specifies one or more decision processes that are deemed appropriate to that problem. These decision processes are called the "feasible set" and represent the methods that remain Situational Variables after a set of seven rules has been applied. The first three of these rules eliminate methods that threaten the quality of the deci.sions, while the last four rules eliminate methods that are likely to jeopardize acceptance of the decision by subordinates. For those who are unfamiliar with the Vroom-Yetton model, let me point out that the decision processes arc described here in a kind of code. AI and All are variants of an autocratic process. In AI the manager solves the problem by himself using whatever information is available to him at that time; in All he obtains any necessary information of a specific nature from his subordinates before making the decision himself. CI and CII are variants of a consultative process. In CI he shares the problem with relevant subordinates individually, getting their ideas and Figure 2 DECISION PROCESS FLOWCHART (FEASIBLE SET) A. B. C. D. E. Does the problem possess a quality requirement? Do I have sufficient ir)formation to make a high-quality decision? Is the problem structured? Is acceptance of the decision by subordinates important (or effective implementation? If I were to make the decision by myself, am t reasonably certain that it would be accepted by my subordinates? F. Do subordinates share the organizational goals to be attained in solving this problem? G. Is conflict among subordinates likely in preferred solutions? suggestions before making the decision; CII is similar, but the consultation takes place within the context of a group meeting. Finally, CII corresponds with Norman Maier's concept of group decision in which the manager's role is that of chairperson of a group meeting aimed at reaching consensus on the action to betaken. The part of the model described so far specifies bow decisions should not be made, not how they should be made. For most problem types, there exist more than one decision process consistent with the rules and therefore contained in the feasible set. We have also been concerned with the consequences of various ways of choosing from these alternatives. There is considerable evidence that the time required to make the decision (defined either as the elapsed time or the number of man-hours needed to make the decision) increases with the intensity of involvement or participation of subordinates. Thus a time-efficient model (which we term Model A) would select the most autocratic alternative within the feasible set, a choice that would be clearly indicated in crisis or emergency situations and in situations in which one seeks to minimize the number of man-hours that enter into making the decision. Of course, time is not the only dimension to include in deciding the degree to 19 which the leader should encourage the participation of his subordinates in decision making. In addition to the possibilities that participation may increase decision quality or its acceptance (considerations that are incorporated into the rules referred to previously), there are also grounds for believing that participation contributes to individual and team development and is likely to result in more informed and responsible behavior by subordinates in the future. Hence Model B, which could be thought of as a time-investment or developmental model, dictates the choice of the most participative process within the feasible set. It is important to note that Models A and B are consistent with the same rules (to protect decision quality and acceptance) but represent extremely different ways of operating within these rules. Model A maximizes a short-run value—time; Model B maximizes a long-run value—development. What is the image of the effective leader portrayed by this normative model .^ He is neither universally autocratic nor universally participative but utilizes either approach in response to the demands of a situation as he perceives them. Above all, he is a flexible leader who has thought through his values and who has a repertoire of skills necessary to execute effectively each of the decision processes. VALIDATING THE MODEL 20 When Philip Yetton and I wrote our book, we had no evidence validating the model other than the consistency of our rules with existing empirical evidence concerning the consequences of alternative approaches. During the past six months. Art Jago and I have been working to remedy this deficiency. We have asked managers, all of whom were unfamiliar with the model, to select two deci- sions that they had made—one that proved to be successful and one that proved to be unsuccessful. Each manager wrote up each decision situation as a case and specified the decision process he used in solving the problem. Later these managers were trained in the problem attributes and went back over each of these two cases, coding each in a manner that would permit the researcher to determine the problem type and the feasible set of methods for that problem type. The data for this study are still coming in. To date, we have written accounts of 46 successful decisions and 42 unsuccessful ones. (It seems that some managers have difficulty in recalling the decisions they made that did not turn out too well!) Figure 3 shows the results available so far. These results clearly support the validity of the model. If the manager's method of dealing with the case corresponded with the model, the probability of the decision's being deemed successful was 65 percent; if the method disagreed with the model, the probability of it's being deemed successful was only 29 percent. It is important to note, however, that behavior that corresponds with the model is no guarantee that the decision will ultimately turn out to be successful—nor is behavior outside the feasible set inevitably associated with ail unsuccessful decision. To create a model of decision processes that completely predicts decision outcomes (that is, which generates 100 percent observations in upper left and lower right cells) is an impossibility. Any fantasies that we may have entertained about having created a model of process that would completely determine decision outcomes have been permanently dashed against the rocks of reality! Insofar as organizations are open systems and decisions within them are made under conditions of risk and uncertainty, it will be impossible to generate complete predictability for a model Figure 3 RELATIONSHIP BETWEEN MODEL ACREEMENT AND DECISION OUTCOME Percent Percent Sue- Unsuccessful cessful Total Method used agrees with feasible set 65 35 100% Method used disagrees with feasible set 29 71 100% such as ours. To be sure, we may be able to use the data from the study I have described to improve the "batting average" of the model, but the limit of success must be less than perfection. IMPLICATIONS FOR TRAINING I would now like to turn to the central issue of this symposium, the use of the model in leadership training. Over the past few years, several thousand managers have received training in the concepts underlying the model. The workshops have ranged from two to over five days in length, and the participants have included admirals, corporation presidents, school superintendents, and senior government officials. I have been personally involved in enough of this training to have learned some important things about what to do and what not to do. And because I believe that there are substantial but understandable misconceptions about how training based on the Vroom and Yetton model works, I would like to describe the things I have learned. It would have been possible to build a training program around the model that was completely cognitive and mechanistic. Participants would be sold on the model and then trained in its use through intensive practice— first on standardized cases and later on real problems drawn from their own experiences. Such an approach would represent a new domain for Taylorism and could even be accomplished through Skinnerian programmed learning. I believe that, at best, this behavioral approach would influence what Argyris calls espoused theories and would not have any long-lasting behavioral effects. Our methods have been much more influenced by Carl Rogers than by B. F. Skinner. We have assumed that behavioral changes require a process of self-discovery and insight by each individual manager. One method of stimulating this process is to provide the participant with a picture of his own leadership style. This picture includes a comparison of his style with that of others, the situational factors that influence his willingness to share his power with others, and similarities and differences between his own "model" and the normative models. In advance of the training program, each participant sits down with a set of cases, each of which depicts a leader confronted with an actual organizational problem. We call these cases "problem sets," and the number of cases in different problem sets ranges from 30 to 54. The common feature in each of the eight or nine problem sets that have been developed is that the cases vary along each of the situational dimensions used in the construction of the normative model. The set is designed so that the variation is systematic and that the effects of each situational attribute on a given manager's choice of decision process can be readily determined. This feature permits the assessment of each of the problem attributes in the decision processes used by a given manager. The manager's task is to select the decision process that comes closest to depicting what he would do in each situation. His re- 21 Victor H. Vroom is fohn G. Searle Professor oj Administrative Sciences and professor of psychology at Yale University. Since 1974, he has been involved in plans jor Yale's new School oj Organization and Management and has served as chairman oj the interim board oj governors of that school. He received his bachelor's and master's degrees jrom McGill University and his Ph.D. degree jrom The University oj Michigan. He was the winner oj the Pord Foundation Doctoral Dissertation Competition at Michigan and held the Ford Foundation's Faculty Fellowship at the University oj Pennsylvania. He has also won the McKinsey Foundation Research Design Competition (1967), been appointed Fulbright Lecturer in the United Kingdom, and received the fames McKeen Cattell Award oj the American Psychological Association (1970), Dr. Vroom is a director of the Connecticut Savings Ban}{^ and of the Connecticut Mental Health Center. He has been active as a consultant to over 30 major corporations, including General Electric, Bell Telephone Laboratories, and Chemical Bank^. He is a jellow oj the American Psychological Association and has served on its council oj representatives. For the past 20 years. Dr. Vroom has been involved in research and teaching on the psychological analysis oj behavior in organizations. His book^ Work and Motivation is regarded as a landmark in the field. He has also authored six other boof(s, most recently Leadership and Decision Making with Philip Yetton. 22 sponses are recorded on a standardized form and processed by computer along with other participants' responses in the same program. Instead of writing about information contained on a printout, I thought that it might be more efficient to let you see what it looks like. The next figure reproduces three of the seven pages of feedback that a manager recently received. Examine the first page o£ the printout shown in Figure 4. Consider A first in that figure. The first row opposite "your frequency" shows the proportion of cases in which the manager indicated he would use each of the five decision processes. The next row (opposite "peer frequency") shows the average use of these processes by the 41 managers constituting his training group. A comparison of these two rows indicates the methods he used more and less frequently than average. The third row shows the distribution of decision processes that would be used by a manager using Model A, the time-efficient model in the 30 cases. The final row shows a distribution for Model B, the developmental or time-investment model. To obtain an overall picture of how participative this manager's responses are in relation to other members of his training group and to Models A and B, it is necessary to assign scale values to each of the five decision processes. The actual numbers used for this purpose are based on research on the relative amounts of participation perceived to result from each process. AI is given a value of 0; All a value of 1; CI a value of 5; CII a value of 8; and Gil a value of 10. With the aid of these scale values a mean score can be computed for the manager, his peers, and both models. These are obtained by multiplying the percentage of times each process is used by its scale value and dividing by 100, These mean scores are shown in B along with the standard deviation (SD), a Figure 4. Page 1 of Printout LEADERSHIP STYLE ANALYSE NAME OR I . D . - - JOHN DOE A-- B - S C A L E D PARTICIPATION SCORES PROPORTrON OF CASES tN WHICH EACH DECISION PROCESS IS USED 01 CII GM YOUR FREOUENCY 43% AI 3% 23% 30% 0% PEER FREOUENCY 25% 14% 19% 27% 15% MODEL A (MINIMIZE PARTICIPATION) MODEL B (MAXIMIZE PARTICIPATION) Alt 40% 13% 3% 23% 20% 0% 0% 0% 40% 60% MEAN YOUR RESPONSES SD 3.60* 3.48 PEER AVERAGE 4.73 3.66 MODEL A 4.17 4.31 MODEL B 9.20 0.96 * - YOUR SCORE IN FIGURE 1 C--FREOUENCY DISTRIBUTION OF SCALE SCORES (MEAN LEVELS OF PARTICIPATION) YOUR MEAN PEER MEAN LOW PARTICIPATION HIGH PARTICIPATION P A R T t C I P A T I V E N E S S measure of dispersion around the mean—that is, an indicator of how much behavior is varied over situations. These mean scores are shown graphically in the figure at the bottom. Each asterisk is the mean score of one of the group members. The symbol X is printed underneath this manager's mean score, the symbol P under the group average, and the symbols A and B show the location on the scale of Models A and B respectively. corresponding problem numbers are shown In the two left-hand columns of D, In the third and fourth columns, the prescriptions of Models A and B arc given, and the fifth column shows the feasible set for that problem type. The last column, marked "your behavior," indicates the manager's responses to each of the cases of the indicated problem type. If there is more than one case of that type, the methods used are shown in the same order as the problem numbers at the left-hand side. D through F in Figure 4 are on the second page of the printout. As we previously mentioned, the normative model identifies 12 problem types corresponding to the terminal nodes of the decision tree shown in Figure 2. There is at least one case within the set of 30 problems that has been designated by the authors and most managers as representative of each type. The problem types and E reports the frequency with which the manager's behavior agreed with the feasible set, with Model A, and with Model B. For comparison purposes, the average rates of agreement for members of the manager's training group arc also presented.