As with all techniques, the use of the Reflective Thinking procedure has both advantages and disadvantages. On the plus side, the model leads group members to explore each proposal methodically. It can also help members make decisions in as optimizing and unbiased a manner as possible. The procedure can do this because it separates "problem analysis," "solution generation," and "solution evaluation." The model also has an air of objectivity because it uses criteria as the major basis for evaluating proposals. This objectivity may help to soften the hard feelings that could result if the group does not accept certain members' "pet" ideas.
The Reflective Thinking model does have some drawbacks. One weakness is that the method assumes that people can reach consensus on criteria. This idea, in turn, presupposes that group members have a common set of values. Such values would include, for instance, what is important and what is not. For example, woe to the Diner's Club if half its members hungrily demanded a lot of food quickly, while the other half preferred to spend time preparing something really good! It may be difficult for members to agree about criteria.
Further, the model assumes that people have the ability to optimize fully. In particular, the method holds that people can think of criteria for judging solutions before they think of the solutions themselves, which is extremely difficult. It is more "natural" to think of criteria while judging proposals. Groups that use the Reflective Thinking procedure can fall into this "natural" process; the problem is that groups may neglect to evaluate certain proposals against all criteria. For example, in discussing its third alternative solution, a group may think of a new criterion that is particularly relevant. The group could go back and judge the first two solutions against this new criterion also. However, a group may neglect to do this. If this happens, the first two proposals have an unfair advantage. As you know, a solution will remain in the running as long as it does not fail to meet a criterion. This means that the fewer criteria that a group uses to evaluate a proposal, the less likely the possibility that the group will reject the idea. Therefore, in this case, it is unfair if the group judges the third proposal against one more criterion than the first two options. This would mean that the third option unfairly has a greater chance of being rejected.
A group can solve this difficulty by conducting a preliminary evaluation of proposals. During this preliminary discussion, the group can uncover new criteria as it examines the proposals. The group can then go back and perform a final evaluation, in which the members judge solutions against all the criteria. The problem with using both a preliminary and a final evaluation period is that it makes the procedure take even longer than usual.
Keeping these drawbacks in mind, it appears that groups should save the Reflective Thinking technique for two basic circumstances: (1) when the problem is very important and the group has a lot of time and patience for solving it and (2) when a standard set of criteria already exists that the group can use to evaluate proposed courses of action.
Table 13.4 summarizes the steps in Reflective Thinking.
TABLE 13.4 Summary of steps in Reflective Thinking.
Step 1. Awareness of the difficulty
A. What is a precise definition of the difficulty?
B. What are the symptoms of the difficulty? How has the problem manifested itself? Whom does it hurt, and how does it hurt them? Under what conditions does it hurt them?
C. How big is the problem? Is it getting worse?
D. What are the implications of the difficulty in the future? What results can be expected if the problem is not solved?
E. What is being done at present to meet the problem? In what ways are these efforts ineffective?
Step 2. Analysis of the problem
A. What causes led to the present difficulty? What conditions exist in the situation that allow the causes to act as they do?
B. Which causes are major and which are secondary?
C. What direction should our approach take? Should it deal with major causes, secondary causes, or symptoms of the problem?
D. Which requirements must a satisfactory solution meet? Do they include any of the following general criteria: the extent to which the proposal would eliminate the difficulty, the extent to which the proposal is feasible in terms of time, person-power, expense, and material? Are any further criteria specific to this difficulty?
E. Must a course of action observe any "boundaries"? Do social customs, institutions, laws and soon place a boundary on the feasibility of possible solutions? Should any of these serve as additional criteria that possible solutions should meet?
Step 3. Proposal of possible solutions
(perform through the use of either brainstorming or the silent generation of ideas)
Step 4. Evaluation of possible solutions
A. To what extent would each proposal meet each criterion for a satisfactory solution?
B. Which proposal best meets the criteria?
Step 5. Implementation of chosen solution
(perform through the repetition of the entire procedure)
As you know, the Reflective Thinking decision procedure demands that group members follow steps that lead to extreme optimization. This is often difficult for members to do. Incrementalism, a satisficing model, is one reaction to this difficulty.
Lindblom (1959), who created incrementalism, observed that "real life" decision makers in industry and government tend to consider only a narrow range of alternative solutions. These alternatives differ by only a small, or "incremental," degree from the status quo. Lindblom did not call this "narrow-mindedness" and criticize these decision makers. Instead, he praised the virtues of their "incremental" procedure. Lindblom's praise was based on his conviction that attempts at optimization are doomed to fail. Why did Lindblom believe this?
First, Lindblom agreed with Simon's argument that optimization is impossible because it is too demanding on our cognitive capabilities. Second, Lindblom observed that optimization requires group members to know a great deal of information about consequences and alternatives. Decision makers will possess this information if they have had previous experiences that are similar to their present situation; however, they will not have the knowledge if the circumstances are new to them. If people attempt to optimize in situations in which information is scarce, they are doomed to failure.
Third, optimization requires group members to agree on underlying values. There must be a consensus. This does not often exist when people are grappling with large, complex problems. Even an individual will have difficulty formulating a consistent set of underlying values to use as a basis for decision making. Fourth, the costs of optimization are prohibitive in resources and person-power. In addition, attempts at optimization are very time-consuming. By the time the group reaches a decision, the solution may no longer be applicable. For example, a plan to save a business from bankruptcy may be very slow in coming. If it is too slow, the business may be gone before the company can implement the plan.
As you can see, Lindblom found many reasons for believing that the process of optimization has many problems. He believed that the "incrementalist" procedure would alleviate all these problems. Decision makers can adopt Lindblom's procedure by using the following process:
1. List only those alternatives that differ "incrementally" from the status quo. These differences are based on the known or expected consequences of the alternatives. In addition, the solutions must be clearly feasible in terms of time, money, and effort to implement.
2. Compare each alternative, in turn, with the status quo and with the other possible solutions. Look for a proposal that has the best immediate consequences. Do not consider long-range goals and ideals.
3. Choose the best alternative through a voting process.
4. If a new problem appears, go through the entire procedure again.
Lindblom hoped that a series of incremental changes of this sort would lead to a favorable outcome and that, consequently, group policy would steadily, if slowly, improve. Lindblom called this process "muddling through."
Incrementalism has some virtues as a decision-making procedure. These virtues, however, are exactly the opposite of what Lindblom believed they would be. Lindblom believed that his technique would be useful in the following situation: (1) the consequences of alternative courses of action are uncertain, (2) the current situation is unacceptable, and (3) the stakes are high. Lindblom was wrong, however. In this type of situation it is best that groups not use his incrementalism process. Instead, groups facing this situation must expend the effort necessary to make an optimal decision.
What type of situation, then, could reveal the virtues of Lindblom's process? His method is useful if (1) members know the consequences of alternative courses of action, (2) the current situation needs only slight adjustments, and (3) the stakes are low. In this type of situation, the incrementalism procedure is a useful way to save a group from the unnecessary detail inherent in optimizing procedure.
In Chapter 12, we discussed the theory that Janis and Mann (1977) proposed regarding how levels of arousal affect decision making. Their theory is useful for a group trying to decide whether or not to use Lindblom's method. As you recall, Janis and Mann proposed a series of questions within their decision-making process. The first two questions were, "Is the present course of action sufficient?" and "Is the most available alternative a sufficient improvement?" These two questions are identical to the incrementalist strategy. If a group answers "yes" to either question, Lindblom's model is adequate for the group. If, however, the answer to both questions is "no," the group should move on to a more optimizing procedure.
In addition, the incrementalism procedure has another virtue. It uses a voting strategy, which is often a good idea; however, a group must be careful when it uses a vote to solve problems. It is true that the optimizing method, calling for a consensus on criteria, is problematic. If serious differences in values exist among group members, reaching a consensus is difficult. Thus, voting is often the only reasonable strategy to resolve group conflicts. There are disadvantages to voting, however. The final vote may ignore the needs of minority groups, and the group cannot claim that it reached its decision using objective criteria.
Thus, a group may need to perform an incremental strategy in response to inherent group conflict. If it does, however, the group needs to be aware of the potential problems when it votes. The group should make concessions to the minority in its decision. It should also ensure that all interested parties are involved in the decision-making process.
The incrementalist procedure may alleviate the problems we have outlined above in regard to optimizing; however, Lindblom's method is badly flawed. Lindblom is correct that optimization is problematic; however, he tries to alleviate its difficulties by wishing them away. For instance, it is true that the members need not worry about a lack of information if they limit their alternatives to those that differ only slightly from the status quo. Is this an adequate solution? It is also true that groups do not have to worry about differences in underlying values if a"winner-takes-all" vote decides the issue. Again, is this a good way to alleviate the problem?
People do not invest much time and effort, nor do they tax their cognitive abilities, if they use Lindblom's incrementalist procedure; however, they also do not have the opportunity to come up with fundamental improvements in current policy. As you can see, Lindblom's procedure may address the problems of optimizing, but it does so in a way that limits decision makers.
Further, Lindblom's procedure appears to be dangerously conservative, despite his claims that it is not. The incrementalist process is useless when a group needs to make a major change in policy. The method also allows voting majorities to maintain the status quo, no matter how reasonable the arguments of minorities. In addition, the model is aimless. Instead of moving toward goals, it moves away from problems. Lindblom sees this as a virtue. We do not see it that way. With no goal in mind, a group can "muddle" its way to disaster. Similarly, a group could make a series of incremental changes in the right direction but still be unable to stop disaster from overtaking it.
RESEARCH ON FORMAL PROCEDURES
Thus far in this chapter we have examined why scientists believe that groups ought to make better decisions when they use formal procedures than when they use a free discussion format. We have described some formal procedures and also discussed how each one's design should fit certain situations and not others.
Except for our examination of brainstorming, however, we have not talked about much of the research on this topic. Many studies have tested the claims that scientists make about formal procedures. For the remainder of this chapter we will present an analysis of this research.
Point for Researchers to Keep in Mind
Before our analysis, however, we must remind our readers of an important point we made at the beginning of this chapter. Formal procedures are in some sense "unnatural." As a result, it can be difficult for groups to follow them. This point has two important implications.
Training and Practice Important
First, people should not expect groups to follow formal procedures correctly without proper training and practice. Members need to get adequate instruction in the procedure and have the opportunity to practice it. Researchers testing the value of formal procedures need to keep this in mind. Their research should be with groups that have sufficient instruction and practice. Groups could follow a procedure poorly because they are not adequately prepared. If this happens, the study will not give the value of the procedure a fair test.
Second, even if group members have sufficient training and practice, they still may not follow the procedure well if they do not care about the decision. Going through an "unnatural" discussion can be difficult if the topic does not seem important. A procedure will not help a group make a good decision if its members merely go through the motions of doing it. Again, the value of formal procedures will not get a fair test in this situation.
Hence, motivation is important. Members will do best when they are making a decision that matters to them. They are more likely to apply themselves to their task and more likely to follow a formal procedure well than they would otherwise. When researchers want to study the value of formal procedures, they need to be aware of this. They should study groups in which members make decisions that affect themselves. Only then can researchers assume that group members are motivated to try to follow procedures well.
Thus, researchers who evaluate discussion procedures should keep in mind that they need to study group members who are adequately prepared and motivated.
In the following pages we will examine research in the area of discussion procedures. To do so, we will ask five questions. As we come to each question, we will discuss the scientists' attempts to answer it. Hence, we will examine each question in turn.
Question 1--Do Discussion Procedures Help Groups Make Better Decisions?
On first glance, it appears that a researcher would have no trouble designing a study to answer this question. A scientist should simply train some groups in a discussion procedure and then give them a decision to make that matters to them. He or she then should form comparable groups and have them use free discussion. Finally, the scientist should determine whether the groups that used a procedure made better decisions than the other groups. Things are not that easy, however. Let us first look at some studies that have attempted to answer this research question. We shall then examine some problems with the studies.
Evidence That Formal Procedures Are Helpful
First, some evidence indicates that groups that follow formal procedures make better decisions than free discussion groups. We shall present this evidence as it appeared from the studies and afterward discuss drawbacks to the findings.
Procedures help accuracy
For example, remember our discussion of the "Lost on the Moon" problem-solving task? We described it in Chapter 2. In this task, group members look at items that can help a stranded crew get back to its spaceship, and they rank the items in order of importance. In our earlier discussion, we said that group rankings for this task tended to be less accurate than the rankings of the most accurate members. Evidence indicates, however, that procedures can perhaps help groups with these kinds of tasks.
In several studies, groups have followed a formal procedure for "Lost on the Moon" and similar "survival" tasks. The groups used the Hall and Watson consensus guidelines. These groups made more accurate rankings than free discussion groups. In addition, they often made rankings that were as accurate as their most accurate member. In other words, Hall and Watson's rules appear to help groups take advantage of their most competent member's knowledge. Apparently, a procedure can enhance the accuracy of a group's solution to a problem.
Procedures help decision quality
Other studies have examined how procedures affect the quality of a group's decision. These studies have shown that groups can improve their decision quality by using a formal procedure. For example, Larson (1969) looked at groups that used three procedures, including Reflective Thinking, to approach decision-making. Larson found that decision quality was higher in these groups than in free discussion groups.
Researchers have found other advantages as far as formal procedures and decision-making tasks are concerned. Delbecq, Van de Ven, and Gustafson (1975) compared groups using NGT with free discussion groups. When they compared them with other groups, the researchers found that groups that use NGT:
1. tend to generate more nonoverlapping ideas (this advantage increases as the size of the groups becomes larger);
2. feel more free to participate, leading to less conformity among group members;
3. are better able to "depersonalize" the discussion and face any conflict; and
4. are less cohesive but more satisfied with task accomplishments.
Procedures help member satisfaction
At the beginning of the chapter, we described Poole's eight reasons for believing that formal procedures are better for group decision making than free discussion. Some of these reasons are relevant to group member satisfaction. For example, formal procedures are thought to help group members feel that their discussion is organized, that they can evaluate the quality of their decision making, and that they are in control of their destiny. Thus it follows that the adoption of formal procedures should increase group member satisfaction. We have already mentioned one study with findings that support this notion (Kramer et al., 1997), and there are several others (for example, Green & Taber, 1980).
Procedures help commitment levels
Finally, evidence also indicates that formal procedures enhance the commitment that members feel toward their decision. White, Dittrich, and Lang (1980) looked at groups of nurse administrators who made decisions about how to perform job-related tasks. Some groups used either NGT or a procedure that was similar to Reflective Thinking. Others used free discussion. When compared with the others, the nurses who used the procedures in their groups were more likely to implement their decisions in their work. This finding implies that formal procedure groups are more committed to their decisions than free discussion groups are.
Problems with the Studies
Lack of training and motivation. Unfortunately, as we have stated, there are several problems with these studies. For one thing, in all studies we mentioned, the participants lacked preparation. They received little instruction and no practice. Also, the question of motivation presents a problem. The White, Dittrich, and Lang (1980) study was the only one in which the decision affected the participants. Thus, we can question the validity of all the studies. Of course, the results were in favor of the formal procedures. Given this, one can imagine that the participants followed the procedures reasonably well.
Lack of explanations for findings. Another problem with these studies is more subtle than those we have mentioned. It is not at all clear why the studies got the results they did. Why did formal procedures result in more accurate rankings and higher quality decisions than the free discussions did?
Possible Explanations for Evidence
Improved group discussion. One possible explanation is that the procedures led group members to have a more structured and well-reasoned discussion than they otherwise would have. This, in turn, produced an improved performance. This explanation focuses on group process. It is an example of the "input-process-output" model of group discussion that we examined in Chapters 1 and 8. The procedures affected group process and thereby helped group output.
None of the researchers that we have discussed, however, looked at group process. Instead, they linked group input directly with group output. The input consisted of the presence or absence of a procedure, and the output was the group's task performance. Therefore, there is no evidence from these studies that the groups that used formal procedures performed their discussions any more skillfully than the free discussion groups did.
Improved individual performances. Further, it is easy to argue that group process is not actually a factor. Did group discussion actually lead to the improved performances of groups that used procedures? For a possible answer, we can turn to ideas from Hewes. As you can recall, in Chapter 8 we discussed Hewes's (1986) arguments concerning the "input-process-output" model. One of his arguments was that group input itself may affect how group members perform. Differences in the input may lead to variations in the way that individual members perform. Individual performances, in turn, may account for the different group output. All this would mean that group process does not matter. Instead, group input is the key to what happens.
How does Hewes's argument relate to formal procedures? The procedures constitute group input. When group members adopt a procedure, they may improve their thinking. This, in turn, means that the group will make a better decision. Evidence indicates that certain ways of thinking can help decision quality. The ability to think reflectively appears to improve the quality of decisions (Sharp & Milliken, 1964). Procedures such as Reflective Thinking may help group members think more reflectively on their own. If so, these procedures would improve decision quality by themselves, as an input variable. Their effect would be independent of the discussion process.