بهره وری کافی نیست: مقایسه گروه طوفان مغزی تعاملی و اسمی در خلق ایده و گزینش

The conclusion that nominal brainstorming groups outperform interactive brainstorming groups has been exclusively based on studies of idea generation. This study tested whether the productivity advantage of nominal groups would also result in better idea selection. Nominal and interactive groups performed a task that involved idea generation and selection. Idea generation and selection were strictly separated for half the groups, but were combined for the other half. Nominal groups generated more ideas than interactive groups, and the ideas generated by nominal groups were more original and less feasible than the ideas generated by interactive groups. However, there were no differences among conditions in quality of the selected ideas. Further, idea selection was not significantly better than chance. This suggests that high productivity in brainstorming is not sufficient to lead to better solutions.

Ever since the publication of Osborn’s (1953) influential book Applied Imagination, many organizations consider group brainstorming as a particularly effective technique for generating large numbers of creative ideas. This popularity persists despite repeated findings that brainstorming groups generate fewer ideas, and fewer good ideas, than individual brainstormers whose ideas are pooled (so-called nominal groups; e.g., Diehl and Stroebe, 1987 and Taylor et al., 1958). In fact, most people continue to believe that group brainstorming yields more ideas than individual brainstorming. This false belief has been dubbed “the illusion of group effectivity” ( Paulus et al., 1993 and Stroebe et al., 1992). As a result of this discrepancy between research findings and everyday beliefs, brainstorming research has focused almost exclusively on productivity (i.e., number of ideas generated) as a dependent variable (Sutton & Hargadon, 1996), studying possible explanations for the productivity loss in interactive brainstorming groups (e.g., Diehl and Stroebe, 1991 and Paulus and Dzindolet, 1993) and ways to minimize productivity loss (e.g., Gallupe, Bastianutti, & Cooper, 1991). Considering that the brainstorming procedure was designed with the specific goal of boosting productivity, this makes sense. However, when brainstorming is regarded as an innovation tool, this focus is too narrow. After all, producing large numbers of ideas is never the ultimate goal of a brainstorming session. Instead, what brainstormers are after is a limited number of good ideas to select for further development and, eventually, implementation (Nijstad & De Dreu, 2002). Thus, for creativity to become innovation, divergent idea generation must be followed by convergent idea selection. Given the neglect of idea selection as a research topic, we agree with Sutton and Hargadon (1996) that “it is premature to conclude that face-to-face brainstorming groups are ineffective” (p. 688). It should first be assessed whether the higher productivity of nominal groups also leads to a higher quality of selected ideas. The primary aim of our study is to answer this question through a comparison of interactive and nominal groups on a task that involves both idea generation and idea selection. Even though idea generation and idea selection are both essential parts of the innovation process, combining them effectively is considered an onerous task. Brainstorming conventions dictate that idea selection should be separated from idea generation as strictly as possible. An important ‘active ingredient’ (Smith, 1998) in the brainstorming procedure is the deferment of judgment: people are thought to generate more ideas when they feel free of evaluation and criticism (e.g., Amabile, 1979 and Hennessey, 1989), so separation of idea generation and selection is essential. Although the brainstorming literature provides some support for this assumption (e.g., Camacho and Paulus, 1995 and Diehl and Stroebe, 1987), it remains an open question whether this task separation has any effects on the quality of selected ideas. Thus, our second aim is to study whether the recommended separation between idea generation and selection improves the quality of selected ideas. In this study, we manipulate this task separation by presenting idea generation and idea selection either as two separate tasks, or as one task consisting of two activities (idea generation and selection).Idea selection in interactive and nominal groups The outcome of idea selection is dependent on two factors: the quality of the available ideas, and the quality of the selection process. Our manipulation of group interaction could have opposing effects on these factors. With regard to idea generation, the superiority of nominal groups is clear. In line with Osborn’s (1953) proposal that quantity breeds quality, research has shown a strong correlation between the total number of ideas generated and the number of good ideas available therein (Diehl & Stroebe, 1987). Nominal groups generate more ideas than interactive groups, and hence have more good ideas to choose from. With regard to the selection process, however, it is less obvious what to expect. On the one hand, various sorts of process loss can lead to suboptimal idea selection in interactive groups. For example, decision making groups often fail to discuss all relevant information and to consider all available alternatives (see Stasser, 1999). Of course, effective selection largely rests on a thorough consideration of available options. If interactive groups do not consider the options at their disposal, they cannot be expected to compensate for their lower productivity. One plausible prediction, therefore, is that the quality of selected ideas will be higher for nominal groups than for interactive groups. On the other hand, several studies have demonstrated that groups can outperform individuals, particularly on intellective tasks. For example, Laughlin and colleagues (Laughlin et al., 1998, Laughlin and Shippy, 1983 and Laughlin et al., 1991) found that groups performed better than individuals on a task in which participants had to decide between hypotheses to account for certain patterns. Although idea selection is not an intellective, but a judgmental task, group discussion clearly can improve the selection process. Thus, one could also predict that interactive groups will overcome their lower productivity, and will select ideas that are at least as good as those selected by nominal groups. Separation of idea generation and selection We expect that presenting idea generation and idea selection as one task will make the evaluative aspect inherent in idea selection more salient during idea generation, and that this will inhibit productivity. This effect should be especially strong in interactive groups, for two reasons: First, only in interactive groups can members be evaluated by others; thus, we expect any manipulation that increases evaluation apprehension to have a stronger effect for interactive than for nominal groups. Second, interactive groups suffer from production blocking: Members of interactive groups have to take turns expressing their ideas, and this interferes with group members’ idea generation and expression (Diehl & Stroebe, 1991). We expect that removing the task separation will exacerbate this production blocking, because group members will devote more time to explaining their ideas, which means that other group members will have to wait longer before they can express their ideas. The presence or absence of a strict task separation could influence idea selection in two ways. If the separation of idea generation and selection enhances productivity, more ideas are available for selection; this in turn could lead to the selection of better ideas as compared to the situation in which generation and selection are not strictly separated. However, because removing the task separation should cause brainstormers to devote more time to idea evaluation, even while still generating ideas, it could also improve the selection process. Therefore, it is possible that even though productivity will be negatively affected by removing the traditional task separation, the quality of the selected ideas will not be negatively affected.

Previous brainstorming research has focused on idea generation, but neglected idea selection. The primary aim of this study was to fill this void by comparing interactive and nominal groups in an experiment that involved both tasks. Our secondary aim was to study whether the strict separation of idea generation and idea selection is as important as is usually assumed; in order to address this question, we removed this task separation for half of our groups. With regard to idea generation, our study replicated the production loss usually encountered in interactive brainstorming groups: nominal groups generated more ideas than interactive groups. This advantage of nominal groups was even more pronounced when the strict separation between idea generation and idea selection was removed. Our interactive groups generated fewer ideas per minute in the one-task condition than in the two-task condition, which was in line with our prediction that removing the task separation would increase evaluation apprehension and production blocking in groups. However, because one-task interactive groups devoted more time to idea generation, they eventually were as productive as the two-task interactive groups. One-task nominal groups generated slightly more ideas per minute than two-task nominal groups. Because they were also more persistent, one-task nominal groups generated more ideas than did two-task nominal groups. Apparently, removing task separation increased the participants’ motivation to generate many ideas, perhaps because it was more salient that eventually they would need many good ideas to choose from. In interactive groups, evaluation apprehension and blocking counteracted higher productivity, but in nominal groups the increased motivation did lead to higher levels of performance. This finding qualifies the often-assumed benefits of task separation: only in interactive groups did task separation lead to a more efficient brainstorming session (in terms of ideas per minute). Although there were differences in the quality of the generated ideas (nominal groups generated ideas that were more original, interactive groups generated ideas that were more feasible), these differences disappeared in the selection stage. It appears that interactive groups indeed managed to overcome their productivity loss by making an effective selection. However, the quality of the selected ideas was only slightly different from the average quality of the available ideas. The ideas selected by interactive groups were only marginally more original, and marginally less feasible, than their average production; the nominal groups did not even show this small difference. These findings suggest that, with regard to creativity, the selection process was hardly more effective than taking a random sample from the available ideas. This interpretation is supported by the pattern of correlations found, which shows that the quality of selected ideas was related to the average quality, and not the quantity, of the available ideas. Interestingly, the fact that our participants appear to be unable to distinguish good from poor ideas is consistent with findings reported by Simonton (2003) in his analysis of creativity in science and literature. Simonton reports that people are not very good at recognizing their best ideas, and that this does not improve over the course of their career. Our participants’ low selection effectiveness notwithstanding, they were generally satisfied with the quality of their selection. This raises the question of what criteria our participants used when selecting ideas. The experimental task was explicitly framed as a creativity task, the purpose of which was to come up with creative ideas, but we did not explicitly instruct participants to use originality and feasibility as their selection criteria; thus, we cannot exclude the possibility that they used other criteria. For example, because the brainstorming topic we used was highly relevant to our participants, it is plausible that they simply selected those ideas that they found most important, rather than the most original or feasible ideas. This view is supported by inspection of the video recordings: many of the group discussions revolved around specific complaints regarding the Psychology curriculum, teachers and classes. To test the hypothesis that our participants would have been more effective in identifying creative ideas, had they been provided with more explicit selection criteria, we conducted a small study: we presented 30 participants with a randomly chosen set of ideas from the first study, and asked them to select the four best ideas from this set. We instructed half of the participants to use specific selection criteria (originality and feasibility); the other participants were simply instructed to select ‘the best’ ideas. However, the selection effectiveness of participants with explicit criteria was not different from that of participants without selection criteria.4 This makes it highly unlikely that the poor selection performance in our study was due to the absence of explicit selection criteria. Nevertheless, because all participants in this follow-up study worked individually, we cannot exclude the possibility that providing interactive groups with explicit selection criteria would enhance selection performance. Another possible explanation of the low selection effectiveness is that the originality and feasibility of the available ideas may have been strongly negatively correlated, making it difficult for participants to select good ideas. We therefore computed the average correlation between originality and feasibility for the generated ideas and for the selected ideas5 (first performing a Fisher r-to-Z transformation on all correlations, and then transforming the average value back to r). The average correlation between originality and feasibility was only marginally significant for the generated ideas (r = −.284, p = .07) and not significant for the selected ideas (r = −.103, p = .60). Thus, while there was a slight negative association of originality and feasibility in the available ideas, this correlation was not very strong. It should be noted that, in our study, nominal group members generated and selected ideas individually. Thus, none of them had access to the total group’s production. This may have contributed to the nominal groups’ failure to profit from their high productivity. However, given the low effectiveness of the selection process, we suspect that pooling the generated ideas before the selection task would not have resulted in a higher quality of selected ideas. If people are not very good at recognizing and selecting their best ideas, it cannot be expected that selecting from the nominal group’s total production will lead to a higher quality choice. Nevertheless, future research should address this issue in more detail. It is possible that a combination of nominal and interactive idea generation and selection would yield optimal results on both tasks. Another interesting possibility is to study whether the use of different perspectives or goals has an influence on idea generation and selection. It is possible that interactive groups do have some benefit to offer in those cases where idea selection is best undertaken from a variety of viewpoints. In general, ways of improving the selection process seem an interesting area for future study, both from an applied and a theoretical perspective. In conclusion, the existing brainstorming literature is very useful to brainstorming professionals who wish to increase brainstorming productivity. However, we argue that, from an applied perspective, brainstormers and researchers who focus on productivity and neglect idea selection are missing the point: getting brainstorming to yield high-quality solutions. If brainstorming is used as a way to come to a limited number of creative ideas or solutions, productivity clearly is not enough. The advantage of having many creative ideas at one’s disposal can easily be undone by a sub-optimal selection process. Instead of simply making groups more productive, it may be more fruitful to make them more effective in all stages of the creative process.