Enhancing Ideational Creativity in Groups
Enhancing Ideational Creativity in Groups
Lessons from Research on Brainstorming
Abstract and Keywords
This chapter reviews the various social and cognitive factors which affect the generation of ideas in groups. It focuses on ways that one can marshal the social and cognitive processes to demonstrate that group ideation can in fact exceed individual ideation in both the quantity and the quality of ideas produced.
Group creativity comes in many forms. It occurs whenever two or more individuals work together to create some new idea, product, or procedure. These individuals can be casual acquaintances or friends who get together to exchange ideas or solve a problem, or they can be members of an organized work group such as a scientific team, a product development team, or a committee. They can be short-term groups that meet for brief periods of time or long-term groups that work together for months or years on a series of projects. In all of these cases the presumption is that some type of collaboration, interaction, or exchange is important or necessary for creativity, innovation, or problem solving.
There are many accounts of the success of such collaborative adventures (Bennis & Biederman, 1997; John-Steiner, 2000). We do not doubt that such reported successes are real. However, there is an extensive literature on group processes relevant to creativity that suggests ways in which groups may often inhibit the creative process while at the same time giving the illusion to group members and even external observers that a high level of creativity is being attained (Paulus, 2000; Sutton & Hargadon, 1996). In fact, for some time, much of the groups literature suggested that groups were generally inimical to creativity and that collaboration should be avoided at all costs. More recently, we have discovered ways in which group interaction can in fact significantly enhance creativity. Some of our suggestions for improving creative idea generation in groups are supported by computer simulations of an associative memory model of individual and group brainstorming. The focus in this chapter is on (p.111) the literature on group brainstorming or idea exchange using the specific rules suggested by Osborn (1957), the factors that limit its effectiveness, and ways to tap the creative potential of groups.
Productivity in Group Brainstorming
The Origins of Brainstorming
Brainstorming as a technique for idea generation has been used in a variety of group contexts. Hindu teachers in India have brainstormed with religious groups for over 400 years (Osborn, 1957), and Walt Disney encouraged artists to brainstorm in the 1920s. However, it was developed as a formal technique and popularized by Osborn, an advertising executive. Osborn's main concern was with increasing creativity in organizations. He felt that one of the main blocks to organizational creativity was premature evaluation of ideas. This would tend to inhibit the generation or presentation of ideas. Therefore, he proposed that individuals be trained to defer judgment of their own and others' ideas during the idea-generation process. He also emphasized that participants should focus on quantity, trying to generate as many ideas as possible. Brainstormers were encouraged to freewheel by stating all ideas that came to mind and to combine and improve on ideas presented in the group. He cited much informal evidence for the efficacy of following these brainstorming “rules,” and his claim has been supported by controlled research. Brainstorming instructions do enhance the generation of ideas in comparison to conditions without such instructions (Parnes & Meadow, 1959).
The Process Loss Problem
Osborn's (1957) most controversial claim, that group brainstorming could be twice as productive as solitary brainstorming, has been repeated often (Prince, 1970; Rawlinson, 1981) and has stimulated much research demonstrating that, in fact, the opposite is true. Most controlled studies have found that solitary brainstorming is much more productive than group brainstorming (Diehl & Stroebe, 1987; Mullen, Johnson, & Salas, 1991). One problem with Osborn's data gathering approach was that he did not use a key control for the group interaction process: collections of individuals who did not generate ideas interactively. That is, to know whether group interaction is beneficial, one needs to compare the number of ideas generated by a group with the number of ideas generated by the same number of individuals brainstorming alone. The ideas generated by the individual brainstormers can then be combined to form a “nominal group” score that can be compared to the number of ideas generated by group brainstormers. For both the interactive group and the nominal group, only unique ideas are counted; that is, ideas that are essentially the same are not counted more than once. Contrary to Osborn's expectations, research using such (p.112) nominal group comparisons found that interactive groups were actually much less productive than nominal groups (Diehl & Stroebe, 1987; Mullen et al., 1991). For example, Paulus and Dzindolet (1993) found that nominal groups of four produced an average of 77 ideas on the Thumbs Problem (“What would be the advantages and disadvantages of having an extra thumb on each hand?”) in a 25-minute session, whereas interactive groups of four averaged only 45 ideas. This lower productivity of interactive groups in comparison to noninteractive controls has been termed “process loss” (Shepperd, 1993; Steiner, 1972) or “production loss” (Diehl & Stroebe, 1987).
Bases for Production Loss
The initially somewhat counterintuitive finding about process loss in brainstorming groups stimulated considerable research to determine causal factors (see Nijstad, Diehl, & Stroebe, this volume). Some research has documented that concern with evaluation from other group members (Camacho & Paulus, 1995), motivation loss due to lack of accountability for one's individual performance (Diehl & Stroebe, 1987), and the competition for speaking time in interactive groups (production blocking; Diehl & Stroebe, 1991) all may be contributing factors. It seems evident that a number of factors inherent in group interaction make it difficult for groups to reach their creative potential. This past research suggests that the obvious solution is to have individuals brainstorm alone. If group interaction is necessary or desired for other reasons, special efforts should be made to minimize evaluation apprehension, motivation losses, and production blocking (e.g., Pinsonneault, Barki, Gallupe, & Hoppen, 1999).
The data clearly do not support an advantage of groups over solitary brainstormers, but it should be noted that Osborn's (1957) prediction of the efficacy of group brainstorming is usually taken out of context. He described effective group brainstorming in the context of a prior session of solitary writing. Subsequent group brainstorming is deemed more effective than individual brainstorming because
in the same length of time, and under proper conditions, the average person can think up about twice as many ideas when working with a group than when working alone. Nevertheless, nearly all have agreed that an alternation between group ideation and individual ideation is desirable, since a combination of these two methods has produced maximum results in almost every case. (pp. 228–229)
Although Osborn does make the well-known twofold superiority of group brainstorming claim, this statement is strongly qualified. Such enhanced productivity requires proper conditions, and the most effective procedure is presumed to involve both individual and group brainstorming. In a later edition, this statement (p.113) about the twofold superiority of group brainstorming is not repeated, but there is still an emphasis on the utility of combining solitary and group brainstorming: “To insure maximum creativity in teamwork, each collaborator should take time out for solitary meditations. By working together, and then alone, and then together, a pair is more likely to achieve the best in creative thinking” (Osborn, 1963, p. 146). Similarly, he stated, “Despite the many virtues of group brainstorming, individual ideation is usually more usable and can be just as productive” (p. 191). He also noted, “The fact is that group brainstorming is recommended solely as a supplement to individual ideation” (p. 142). Thus, in his 1963 edition, Osborn generally propounds the superiority of individual brainstorming, possibly in light of laboratory research indicating that brainstorming groups were not very productive (Taylor, Berry, & Block, 1958).
Osborn (1957, 1963) and other practitioners have suggested a considerable number of conditions as being necessary for optimal productivity in group brainstorming (Grossman, Rodgers, & Moore, 1989; Rawlinson, 1981). Groups should be trained for effective brainstorming and should employ facilitators or specially trained members that keep the group on task and highly motivated (Osborn, 1957, p. 235). Another important consideration is that groups should be heterogeneous. Groups in which individuals contribute a diversity of perspectives or knowledge bases allow for greater opportunities for creative combinations of ideas. Pairs of brainstormers who are intellectually compatible may be particularly effective (Osborn, 1963, p. 144). Both individual and group brainstorming are presumed to benefit from preliminary writing sessions, quotas or deadlines, brief breaks, and the use of specific, simple, and subdivided problems (Osborn, 1957, 1963). This chapter evaluates these prescriptions in light of recent data and theory.
Osborn (1957, 1963) also provided theoretical bases for the effectiveness of group brainstorming. Brainstorming in groups allows for “social facilitation” of activity levels. Highly productive members may stimulate others to high levels of productivity. This could involve a simple matching process (Brown & Paulus, 1996; Paulus & Dzindolet, 1993) or a degree of competitive rivalry to see who can generate the most ideas. Groups also allow for social reinforcement of idea generation. Group members may provide social rewards in the form of approval, agreement, or repetition of the same or similar ideas. This may increase motivation for idea generation. Finally, group members can cognitively stimulate mutually related ideas in each other. Thus, Osborn provided a compelling basis for future systematic research and practice on group ideation. We summarize this research, the theoretical perspectives, and implications for practice.
A Social Information Processing Model of Group Brainstorming and Creativity
In examining the group brainstorming process over the past 13 years, we have discovered that this process is influenced by a number of social and cognitive (p.114) processes (Paulus, Brown, & Ortega, 1999; Paulus, Larey, & Dzindolet, 2000). We feel that an understanding of these interrelated processes can help explain the occurrence of both production losses and gains in groups. When people are involved in sharing ideas in groups, this sharing process provides information that can affect both the motivation and the ability to generate ideas. The idea-sharing process will inevitably vary in number and quality of ideas exchanged. When group members are exposed to partners who share many ideas, they may be motivated to perform at a similar high level, and the shared ideas may stimulate group members to think of additional ideas or categories of ideas. In contrast, unproductive partners may lower motivation to generate ideas and will provide little stimulation for new ideas. Of course, there are limits in the ability of individuals to process information as they are attempting to generate their own ideas. Highly productive partners may limit others' opportunities to share ideas, so there is likely to be an optimum pace of information sharing that provides both for high levels of motivation and cognitive stimulation without overwhelming the cognitive capacities of the group members.
Motivation and cognitive stimulation can also be affected by factors outside the group. Group motivation can be affected by intra- or intergroup competition, task structure, and facilitators or group leaders. Cognitive stimulation can be influenced by task structure, the mode of idea sharing (e.g., oral vs. written), and ideas or category information presented by external sources such as facilitators.
Our theoretical focus on the combined social, motivational, and cognitive factors in group performance is also reflected in some prior theoretical models. Baron (1986) has developed distraction-conflict theory of social facilitation to account for the effects of social presence on task performance. In performing tasks in the presence of others, individuals experience conflict between their desire to devote cognitive resources to the task at hand and to attend to the social cues provided by observers or coactors. This attentional conflict can hinder performance on complex tasks. However, because the attentional conflict increases arousal levels, performance of simple or well-learned tasks may be facilitated. Furthermore, the degree of attention to others is affected by principles of social comparison (Festinger, 1954). One is most likely to be concerned about attending to individuals who have some implication for one's self-evaluation (e.g., working on the same task; Sanders, Baron, & Moore, 1978). As we discuss later in the chapter, where an individual focuses attention during a brainstorming task is extremely important; to the extent that attention is drawn away from the idea-generation process itself or drawn away from the ideas being suggested by other group members, the productivity of individual group members, and thus ultimately of the group as a whole, will suffer. We also highlight the role of social comparison processes in the idea-generation process.
Paulus (1983) developed the cognitive-motivational model of group task performance to account for both social facilitation and social loafing effects in task performance. Social factors such as group size and evaluation were assumed to affect motivation, arousal, and task-irrelevant processes (e.g., distracting (p.115) thoughts/anxiety). As with other models of social facilitation, the specific outcomes were dependent on task complexity and characteristics of the social context such as positive or negative social consequences (Geen 1989; Zajonc, 1980). The models by Baron (1986) and Paulus have highlighted the importance of both motivational and cognitive factors in group productivity. We have continued to emphasize the importance of motivational and cognitive stimulation factors in group brainstorming (Brown, Tumeo, Larey, & Paulus, 1998; Paulus, Dugosh, Dzindolet, Coskun, & Putman, 2002).
Although in practice, the differential influence of cognitive and social/ motivational factors may be difficult to disentangle, we discuss them separately because they involve unique underlying processes. Our categorization of social and cognitive factors may also be useful for practitioners as they attempt to develop procedures to optimize group creativity. We first discuss the role of social comparison and related motivational processes. Then we outline the cognitive processes that are involved in the group ideation process.
In discussing motivation, the distinction is often made between intrinsic and extrinsic motivation. Intrinsic motivation is “self-based” and is often seen as critical to creative achievements (Amabile, 1996; Hennessey, this volume). Extrinsic motivation is based on external and social pressures that motivate individuals to high levels of performance to attain rewards or approval. Although intrinsic motivation is a critical factor in the maintenance of creative efforts over a long period of time, in short-term settings like those involved in brainstorming sessions, extrinsic motivational factors may play an important role.
In short-term group brainstorming sessions, there are a number of sources of motivational cues. The experimenter's instructions may communicate the degree of importance of the task or provide other motivational cues. The experimenter or an assistant can serve as a facilitator to guide or motivate the brainstormers during a session. Brainstormers can be provided with high performance standards or comparison information about other successful individuals or groups. Task cues may provide brainstormers with information about what is expected of them. Indicators of the length of the session or breaking sessions into subsessions may affect the motivational level of brainstormers. The extent to which motivation is affected by information from others has been addressed by social comparison theory.
Social Comparison Processes
In his classic papers on social influence processes, Festinger (1954) proposed that we have a drive to compare our opinions and abilities to those of others. This presumes that we have some degree of uncertainty about our opinions and abilities and want to use the comparison process to reduce that uncertainty. Another presumption is that we would like to resolve that uncertainty in a positive (p.116) manner. That is, we would like to discover that our opinions or beliefs are confirmed by others, so we tend to seek out for comparison those who are similar to us in other areas (e.g., Goethals & Darley, 1987). In regard to abilities, we might compare ourselves to those we believe might be slightly below us in ability (downward comparison) so that we can enhance our self-esteem (Gibbons & Gerrard, 1991; Turner, 1978). Alternatively, we might compare ourselves to those we perceive as slightly better than ourselves (Gibbons, Blanton, Gerrard, Buunk, & Eggleston, 2000); in that case, if we turn out to be better than they (e.g., in a game of tennis), we receive an extra boost in our self-esteem. It also gives us a reasonable goal for which to strive in terms of self-improvement. If we do not measure up at a specific time, that of course would be consistent with the fact that the other person is supposed to be slightly better and this would minimize the negative impact on our self-esteem.
In the idea-generation paradigm, social comparison processes can occur in a number of ways. First, group members can monitor the rate of idea generation and compare their own rate with that of other group members. Second, they can monitor and compare the quality of the ideas being shared. This monitoring process can have a number of consequences. It can affect the evaluations of self and group performance. For example, group members may develop an illusion that they are doing well because they are performing similarly to other group members (Paulus, Dzindolet, Poletes, & Camacho, 1993). The monitoring process can also have motivational consequences. Participants may become motivated to compete with each other in producing a high number of ideas or good ideas. Alternatively, they may free-ride on the efforts of the most productive group members by letting them do most of the talking. However, the initially more productive group members may decide that they are doing a disproportionate share of the work and adjust their performance in the direction of the low performers (downward matching). That is, they may decide not to play the “sucker” (Kerr & Bruun, 1983). Each of these types of processes has been observed in brainstorming groups.
Upward versus Downward Comparison
What determines whether upward or downward comparisons will predominate? We presume that if the setting emphasizes accountability and relative performance, social sharing may induce competition. For example, when group members are provided periodic comparison information during the course of a brainstorming session, they demonstrate enhanced performance (Paulus, Larey, Putman, Leggett, & Roland, 1996). In contrast, when group members feel that their contributions are anonymous or cannot be evaluated, there is a tendency of members to show reduced motivation or performance in groups (Diehl & Stroebe, 1987). We often find that under such conditions, individuals will also tend to make downward comparisons (Dugosh & Paulus, 2001; Paulus & Dzindolet, 1993).
So, social influence processes can affect brainstorming groups in a variety of ways. They can demonstrate process gains because of increased competitive (p.117) motivation induced by the interaction process. This may involve mutual matching of high performance standards or levels. On the other hand, they can demonstrate process losses because of the opportunity to hide one's performance in the group or free-ride on the efforts of more productive members. Whether upward or downward processes predominate depends on a number of group characteristics.
For groups to function in an upward fashion it may be important for them to have a group culture that values high standards and performance (Gammage, Carron, & Estabrooks, 2001). This may not exist in short-term laboratory groups unless special procedures are used to induce such a culture. Work groups or teams that have appropriate leadership and strong commitment to shared goals may function quite differently than temporary ad hoc groups. Another factor that may be important is some level of intragroup trust. If group members trust each others' motives, they may compensate for group members who are not performing at a high level (Williams & Karau, 1991).
Social/Motivational Bases for Enhancing Brainstorming
The preceding discussion suggests that social comparison within and between groups can have a strong impact on task motivation. Under the right conditions groups can be disposed toward upward comparison and higher performance levels. However, there are many external factors that also can be influential in producing an upward comparison process or high levels of task motivation. Certainly, groups might be motivated by external rewards such as monetary incentives, grades, and public acclaim. We focus on four factors that have been examined for group brainstorming and performance: comparison information, facilitators, task rules, and leaders.
When individuals are provided comparison information in a context in which goal achievement is valued, upward comparison information processes should be invoked. For example, when individual or group brainstormers are provided high comparison standards, they increase the number of ideas generated (Paulus & Dzindolet, 1993). In this study, participants were given performance expectations that were about twice as high as typical performance. As a result, performance of both nominal and interactive groups increased by about 40%. In a similar vein, when groups or individuals discover they have performed more poorly than other individuals or groups in a brainstorming session, they demonstrate increased performance in a subsequent brainstorming session (Coskun, 2000). In conditions where no feedback or positive feedback was provided, no such increase was observed. Several studies in which participants exchange ideas using computers have found that simple feedback about performance of other brainstormers increases brainstorming performance (Paulus et al., 1996; Roy, Gauvin, & Limayem, 1996; Shepherd, Briggs, Reinig, Yen, & Nunamaker, 1995–1996). (p.118) These findings are consistent with other research on the benefits of goals and competition on task performance (Stanne, Johnson, & Johnson, 1999). However, it should be noted that these types of manipulations do not reduce the performance gap between groups and individuals, as both appear to benefit similarly from such manipulations.
Although Osborn's (1957, 1963) rules are useful in increasing group productivity, they do not effectively address the variety of problems groups encounter. One often discussed problem is that some individuals dominate the interaction process and contribute most of the ideas (Bonito & Hollingshead, 1997), so it is important to encourage participation by all group members. Although there is an emphasis on quantity of ideas, group members often take time that could be used to generate new ideas to elaborate or explain their ideas. There are also often periods of inactivity when no one is presenting ideas. If these periods are long enough, they may be seen as a cue that the group has exhausted its idea store and that their task is finished (see Nijstad et al., this volume). Groups also have a tendency to generate ideas within specific categories of a problem and then go on to other categories; typically, they move on before they have exhausted the available ideas in a category. This could be a significant problem because some of the best ideas may be ones that are not immediately accessible (Brown et al., 1998).
When brainstorming occurs in organizations, groups typically are guided by a facilitator who ensures that groups avoid the above pitfalls (Grossman et al., 1989; Sutton & Hargadon, 1996). A number of recent studies have actually examined the potential benefits of using facilitators. A study by Offner, Kramer, and Winter (1996) used a condition in which trained facilitators helped groups adhere to Osborn's guidelines, prevented group members from engaging in irrelevant or disruptive discussions, and motivated continued performance. They found that groups with a facilitator performed as well as a nominal group. They replicated this effect in two additional experiments (Kramer, Fleming, & Mannis, 2001). The ratings of the participants and observation of effective facilitators suggested that the beneficial effects of facilitators were related either to their motivational effects or their ability to manage the interaction process effectively (e.g., limit interruptions and evaluation).
Oxley, Dzindolet, and Paulus (1996) conducted a similar study in which facilitators were provided with a specific set of guidelines based on a review of the facilitator literature. These included keeping group members focused on the task, not letting them tell stories, not letting them elaborate unnecessarily on expressed ideas, encouraging additional ideation when no one was talking, encouraging nonparticipants to contribute ideas, and reminding group members not to criticize. Group brainstormers in a condition with highly trained facilitators generated as many ideas as nominal (p.119) groups and more ideas than nominal groups toward the end of the brainstorming session. Individuals in nominal groups were not provided with facilitators; thus, it remains to be determined whether facilitated interactive groups can outperform facilitated nominal groups. This question is partially addressed in the next section, where the effect of additional brainstorming guidelines is examined for both interactive and nominal groups. However, it is clear that group brainstormers will benefit from the help of trained facilitators.
Additional Brainstorming Guidelines or Rules
Osborn's (1957, 1963) brainstorming rules were designed to help creative groups function more effectively. Although these rules do seem to be helpful (Parnes & Meadows, 1959), it appears that facilitator guidance can have additional positive effects. One critical feature of facilitation is that groups are provided some additional guidelines. An interesting question is whether just providing groups with such additional guidelines will facilitate their performance. That is, if a few guidelines are helpful, maybe additional guidelines are even more helpful. This issue has been addressed in several studies (Putman, 1998, 2001). Putman provided some groups and individuals with regular brainstorming instructions and others with additional guidelines based on those used by Oxley et al. (1996). The additional instructions provided were as follows: (1) Do not tell stories or explain ideas; (2) when no one is saying ideas, restate the problem and encourage one another to generate more ideas; (3) encourage those who are not talking to make a contribution; (4) suggest that participants reconsider previous categories when they are not generating many more new ideas. To follow the instructions, each participant would have to take an active role in monitoring and guiding the group process. When conditions with the additional rules were compared to those that used only Osborn's rules, there was about a 40% increase in number of ideas generated. Somewhat unexpectedly, this increase occurred for both the group and individual brainstorming conditions, and the presence of an experimenter to remind group members to comply with the rules did not have any additional benefit. So it is evident that providing brainstormers with additional guidelines greatly increases their productivity. At this point, it is not clear why this is the case. The additional guidelines could be increasing the efficiency of the brainstorming process, or they could serve to provide participants additional motivation. Putman (2001) found some support for the efficiency hypothesis in that individuals and groups who received additional instructions used fewer words to express their ideas. We have probably just scratched the surface of discovering useful guidelines and the reasons for their efficacy.
All of the key studies on group brainstorming have used leaderless groups. Informal leaders may emerge, but no formal leaders are appointed to motivate and guide the group process. In most real-world settings, groups or teams often (p.120) have someone whose task is to lead the group toward its goals. These leaders may help set the agenda, guide the interaction process, and maintain a reward or compensation structure. Unlike facilitators, leaders are members of the group and may contribute to the group product. They may not have any special training in group dynamics, but it is hoped that they will have some skill in managing the group process.
A major concern of the leadership literature has been the impact of different styles of leadership. The effects of different leadership styles depend on such factors as the type of task, group type, and the relationship between the group and the leader (Bass, 1998; Chemers, 2001). Although directive leaders who are relatively authoritarian can be useful in motivating groups on simple tasks (Fiedler, 1967), for tasks that require creativity and self-motivation a leadership style that enhances intrinsic motivation and self-confidence may be optimal (Burpitt & Bigoness, 1997; Manz & Sims, 1987). Two styles may be useful in this latter type of setting. Transactional leadership is concerned with setting goals and providing feedback and reward for performance (Bass, 1985). This type of leadership may motivate group members to work hard to reach the implicit or explicit goals. Transformational leadership may be particularly helpful (Avolio & Bass, 1998; Bass & Avolio, 1994) for tasks that require a high level of intrinsic motivation. These types of leaders are sensitive to individual differences, focus on novel perspectives and approaches, and attempt to inspire group members to attain their collective goal in a cooperative manner. Such leaders can increase the confidence or motivation level of the group. This has been termed group potency (Guzzo, Yost, Campbell, & Shea, 1993) or collective efficacy (Bandura, 2000). Both transactional and transformational leadership have been shown to have positive effects on performance (Lowe, Kroeck, & Sivasubramaniam, 1996).
Leadership style has not been examined in conventional brainstorming groups. However, there are several studies of leadership in groups using a computer-based group decision support system. One study examined the effects of leadership style on various phases of the creative process using this paradigm (Sosik, Avolio, & Kahai, 1997). In the first phase that involved group brainstorming, transactional leadership was most effective in increasing ideational productivity. A second, writing phase involved writing a group report on the ideas generated. Transformational leadership had the most positive effect on the quality of the report produced in this phase. This positive effect of transformational leadership on electronic brainstorming was replicated in a second study (Sosik, Kahai, & Avolio, 1998), but in a third study both goal setting (transactional) and inspiration (transformational) were related to ideational creativity (Sosik, Avolio, & Kahai, 1998).
Thus, both transactional and transformational leadership may be useful in creative groups. Transactional leadership may be particularly useful for the production or generation of ideas, and transformational leadership may provide the motivation to persist in evaluating these ideas and possibly implementing them. The most effective leader may thus be one who combines both qualities (Avolio & Bass, 1988).
(p.121) Cognitive Bases for Ideational Creativity
One of the reasons Osborn (1957, 1963) believed idea groups would be highly creative is that he assumed there would be much stimulation of mutual associations. Intuitively, the cognitive benefits of brainstorming in a group seem clear: people believe they come up with ideas in a group that they would not have thought of on their own. This is one reason for the popularity of group idea generation in business and industry. The intuition that groups might facilitate (or “prime”) their members to think thoughts they might not have had in the context of solitary brainstorming is reminiscent of the notion from cognitive psychology that certain ideas or memories are more accessible than others (Tulving & Pearlstone, 1966). The concepts we have stored in long-term memory can be thought of as being connected in a lattice or network in such a way that related concepts are more strongly connected and thus more likely to activate each other (Collins & Loftus, 1975). Thus, concepts that are more closely connected to those that are currently active should be more accessible than those that are less strongly connected to current ideas. This way of representing the idea-generation processes also implies that it is situation- or context-dependent: which ideas are currently accessible depends on what is currently active in working memory.
Semantic Networks and an Associative Memory Model of Group Brainstorming
The notion that our knowledge is stored in some sort of semantic network is now standard and the evidence for it is quite strong.1 Preceding a word with a conceptually related word facilitates the processing of the second word (semantic priming; Meyer & Schvaneveldt, 1971; Neely, 1991); lists of conceptually related items are recalled better than unrelated items, and related items tend to be grouped together during recall even if they were not presented together (clustering; Bousfield, 1953; Mandler, 1967); and often, giving people the category labels for items to be recalled facilitates recall (category cuing; Tulving & Pearlstone, 1966). Clearly, the retrieval of relevant information from one's long-term semantic memory is an important part of the process of group and individual brainstorming: you can't effectively brainstorm on a topic you know nothing about! However, it is recognized that creative idea generation that involves “novel combinations” of existing ideas may not be directly accounted for as the result of associative priming in a semantic network.
To use the semantic network representation as a basis for understanding group brainstorming, many details need to be specified so that quantitative predictions can be made. Because one goal of a model of group idea generation is to account for how groups of four, six, eight, or more interact, it would be unwieldy to explicitly represent four, six, eight, or more semantic networks and the interactions among them. Our approach is to represent a brainstormer's (p.122) knowledge of a given task or problem as a matrix of categories. Each entry in this matrix represents the probability of generating one's next idea from the same category as the previous idea or from a different category. For example, for the University Problem commonly used in our laboratory, participants are instructed to come up with ideas about how to improve their university. For this task, an individual who is currently generating ideas on the topic of classes is more likely to continue on that topic, or a closely related topic such as exams, than to switch to a less related topic such as parking. The probability of staying with the category “classes” is thus higher than the probability of switching from “classes” to “exams,” which would be higher than switching from “classes” to “parking.” The matrix representation also includes a category called the “null category,” which represents the probability of coming up with no idea during a given time interval. As a brainstorming session progresses and new ideas become harder to generate, the probabilities in the null category increase.2 An example of a category matrix like those used in our simulations is shown in Figure 6.1.
A number of individual differences relevant to brainstorming are captured by the matrix framework. Fluency, or the amount of knowledge one has about the brainstorming problem and its categories, is represented as higher probabilities in the main body of the matrix relative to the null category (higher initial probabilities in the null category, which represents the likelihood of coming up with no relevant idea, imply that the brainstormer has less knowledge of a particular category). Convergent and divergent thinking styles also fit nicely into the framework. A convergent thinker is likely to stick with a category and explore it more deeply before moving on to generate ideas from other categories. Thus, a convergent thinker is represented by a matrix with relatively high within-category transition probabilities (the diagonal entries of the matrix). On the other hand, divergent thinkers are more likely to skip around among categories and so are represented by matrices with somewhat lower within-category transition probabilities (and correspondingly higher between-category transition probabilities, represented by the off-diagonal entries of the matrix).3
The property of the cognitive network that is perhaps most crucial to determining the effectiveness of group brainstorming is category accessibility. Relevant categories of ideas that have relatively weak connections to other categories will not generally be explored in isolation. These categories are the ones that require input from others to spark the generation of ideas. For example, a student who lives on campus in a dormitory may be less likely to generate ideas about parking when brainstorming on the University Problem. But if a student who commutes from off campus mentions parking, the dorm dweller may be able to come up with a few thoughts on the matter, perhaps recalling the parking difficulties his or her parents had when they visited campus. In the matrix framework, low-accessible categories are represented by low probabilities in the columns of the (p.123) (p.124) matrix, which determine the likelihood of entering a given category from the other categories relevant to the task. Simulations show that presenting a brainstormer with ideas from low-accessible categories not only increases the number of ideas generated from those categories, but increases the total number of ideas generated overall, thus making the individual a more productive brainstormer (Sherwood, 1998).
This prediction was supported in a study by Leggett (1997). She employed nominal brainstorming groups to evaluate cognitive stimulation in the absence of the negative social influences of others. Leggett played audiotapes containing ideas from either high- or low-frequency categories from the Thumbs Problem (“What would be the advantages and disadvantages of having an extra thumb on each hand?”). Category frequency was determined by examining protocols from participants in previous studies using the Thumbs Problem. A high-frequency category is considered to represent greater accessibility than a low-frequency category. Individuals who were primed in high-frequency categories obtained less benefit than those who were primed in low-frequency categories. In other words, priming categories that are already likely to be utilized did not enhance performance as much as priming categories unlikely to be utilized on one's own. This suggests that presenting primes from low-accessible categories of ideas can increase total idea generation by activating knowledge that would have gone untapped. Leggett also found that brainstormers given six primes from each category generated more ideas than those given only three primes per category, a result also predicted by our associative memory model.
In a related study, Dugosh, Paulus, Roland, and Yang (2000) primed individual brainstormers with ideas from an audiotape in a more ecologically valid way by also including “filler”: comments often made by interactive brainstormers that are irrelevant to the task at hand (e.g., someone might remark “I hated typing class” if typing is mentioned during a brainstorming session on the Thumbs Problem). Providing individuals with external primes increased productivity as expected, again illustrating the potential benefits of cognitive stimulation; however, adding extraneous comments reduced the benefits of priming, illustrating one of the many ways in which group interactions work to subvert the potential benefits of working in a group. This result is predicted by our associative memory model because the irrelevant information causes fewer relevant categories to be active in short-term memory. Ultimately, the model also predicts that when the degree of extraneous information is high enough it will completely overcome the benefits of being primed by relevant ideas; that is, individuals who are primed with relevant ideas can still end up being less productive than an individual who is neither primed nor distracted.
Coskun, Paulus, Brown, and Sherwood (2000) showed that providing external primes also improves the performance of interactive brainstorming groups. However, because providing external primes improves the performance of individual (p.125) brainstormers as well, this technique does not lead to a reduction in the productivity gap. Coskun et al. further demonstrated that the manner in which external primes are presented is crucial. Presenting all of the primes simultaneously is much less effective than presenting them sequentially, that is, one at a time every few minutes throughout the brainstorming session.
Although part of the explanation for the superiority of sequential presentation may be motivational (i.e., perhaps sequential presentation has a “pacing” effect), the associative memory model indicates that the limited capacity of short-term memory is an important factor. Short-term memory can be thought of as the subset of recently activated categories whose activation decays over time but that are still active enough to influence the idea-generation process (cf. Anderson, 1995; Cowan, 1995). In the model, a short-term memory parameter simply assigns a strength to categories from which ideas were most recently generated, and that strength decays fairly rapidly. Simulations demonstrate that the strength of short-term memory does not influence performance when a brainstormer is presented with regularly spaced sequential primes, but that high short-term memory strengths are required if simultaneous priming is to produce a significant benefit over a no-priming control condition. This implies that the effectiveness of sequential primes lies in the fact that they reduce the cognitive load of having to keep track of past ideas. The external primes act as a reminder of the relevant categories, possibly freeing the brainstormer to devote cognitive resources to other aspects of the interactive (and individual) idea-generation process.
At this point it should be noted how the associative memory model accounts for group performance, given that the matrix formalism itself is simply a representation of an individual brainstormer's knowledge and performance. It is the process of attention by which individual brainstormers are assumed to be linked into an interactive group. Individuals will be influenced by other group members to the extent that they are paying attention to each other's ideas. In the probabilistic framework of the model, attention is represented as the probability that an individual group member uses the current speaker's idea as the basis for generating his or her next idea (as opposed to simply continuing his or her own internal train of thought). This allows for possible differences in the degree to which individuals attend to their fellow group members.
Simulations predict that, in general, the more attention one pays to one's fellow group members, the better the performance of the group. Conversely, the more one's attention is distracted from the ideas of others by concern for the social aspects of group brainstorming, the more the performance of the group will decline. In particular, the more one attends to fellow brainstormers, the more one is likely to be primed to consider ideas from one's own low-accessible categories. In fact, the model predicts that, in general, if it were not for production blocking, the number of ideas generated by each group member would increase (p.126) (at least up to a point) as group size increases (Brown et al., 1998). This reflects the results mentioned earlier from Leggett (1997; Dugosh et al., 2000), which indicate that providing brainstormers with external primes, especially primes from low-accessible categories, improves brainstorming productivity. Clearly, there are potential benefits to group idea generation.
One reasonable way to test the effects of attention on brainstorming performance is to instruct brainstormers that at the end of the brainstorming session they will be asked to recall the ideas they were primed with or the ideas generated by the group. There is some evidence that these memory instructions serve to improve brainstorming productivity (Dugosh et al., 2000). Interestingly, however, the effectiveness of memory instructions appears to be mixed. When participants listened to audiotapes or exchanged ideas by computer, instructions to memorize facilitated idea generation during the exposure session and afterwards. Apparently, the memorization instructions increased the extent to which participants attended to the ideas presented and in turn led to additional associations. Without memorization instructions, participants may be more likely to focus on the generation of their own ideas and, to some extent, ignore the ideas that are being presented simultaneously by others. However, when students were asked to exchange written ideas in a round-robin format (Paulus & Yang, 2000), memory instructions inhibited performance. In this paradigm, students were asked to read the ideas as they were passed from one person to the other. Because the instructions to read may have already ensured that participants attended to the ideas of others even while they were in the process of generating their own, instructions to memorize may simply have added an unnecessary extra demand that may have impeded the brainstorming effort (i.e., cognitive overload).
At this point, it is fair to say that although the group productivity gap is persistent in laboratory studies and a number of inhibitory factors have been well-documented, it is clear that under the proper conditions, group brainstorming should provide benefits that in general cannot be realized by solitary brainstormers. Because group interactions can lead to ideas that may not have been generated by solitary brainstormers (ideas from low-accessible categories), it is important to recognize that under the proper conditions, group brainstorming should prove beneficial regardless of whether the productivity gap between groups and individuals (which is measured in terms of quantity of ideas generated) can be overcome.
Cognitive Bases for Enhancing Group Brainstorming
When trying to create circumstances that optimize group performance, the goal is to maximize the benefits of cognitive facilitation (interactive priming) while at the same time minimizing the many inhibitory processes that serve to reduce group productivity. In a nutshell, this statement provides us with a formula for optimizing group performance. Our previous discussion suggests that factors that increase attention to novel ideas and enhance the processing of such ideas will increase the creativity output of groups. We now consider three additional (p.127) methods that appear quite promising for theoretical reasons and that have garnered some empirical support: creating breaks in the brainstorming session, combining group and solitary brainstorming, and having group brainstormers interact by writing instead of speaking (“brainwriting”). Electronic brainstorming, or the use of networked computers on which individuals type their ideas and read the ideas of others, is another potential way to overcome many of the inhibitory factors associated with face-to-face brainstorming. This approach is discussed by Dennis and Williams (this volume).
Mitchell (1998) and Horn (1993; see also Brown & Paulus, 1996) both demonstrated that having individuals take a short break (2–5 minutes) halfway through a 20-minute brainstorming session gives rise to an increase in productivity following the break compared to individuals brainstorming continuously without a break. Mitchell also varied the activities in which brainstormers were engaged during a break from orally brainstorming on the Thumbs Problem. Brainstormers who were asked to continue thinking about the task by writing down any ideas that came to mind during the break and brainstormers who had the brainstorming problem and instructions reread to them during the break both showed an increase in performance following the break relative to brainstormers who continued on without a break. Brainstormers who performed a verbal fluency task (“Think of an object that begins with each letter of the alphabet starting with the letter A”) did not perform significantly better than the no-break controls in the period following the break. Brainstormers who took a break but who were not given any specific instructions or task during the break fell in between the two extremes.
Thus, the specific mental activity in which a brainstormer is engaged during a break is important: the above data clearly indicate that the contents of short-term memory during a break affect an individual's postbreak brainstorming performance. If the activity performed during the break does not allow the task-relevant ideas and concepts to remain active in short-term memory, then the relevant categories will have to be reactivated or “reloaded” following the break. (Low-accessible categories, which, by definition, are less likely to be reactivated, would be most strongly affected by short-term memory interference during a break.) This effect is demonstrated in model simulations: when the current contents of short-term memory are reduced or eliminated during a break, the postbreak performance suffers relative to the condition where the contents of short-term memory remain activated throughout the break.
Smith (1995, this volume) has discussed in detail the importance of brief breaks in overcoming mental blocks of previous experiences. That is, one may become fixated on a particular approach or category during some point in the idea-generation process. Brief breaks or changes to a novel context can provide a stimulus to take a different approach to the problem. In brainstorming settings, brief breaks may lead individuals to switch their ideation to a new domain (p.128) or category from those previously considered. In this way, breaks can be seen as means of overcoming cognitive fixation on a limited range of categories.
Individual and Group Brainstorming
If we take the formula stated above at face value, we should look for some way to literally combine group and individual brainstorming. Of course, a person cannot be in two places at once and therefore cannot brainstorm alone while at the same time brainstorming in a group. But one can alternate group and solitary idea-generation sessions (as Osborn, 1957, 1963, suggested). Preliminary data from one of our laboratories (Leggett, Putman, Roland, & Paulus, 1996) make it clear that brainstorming in a group prior to brainstorming alone on the same topic produces more ideas over the course of the two sessions than does brainstorming alone in the first session and then brainstorming in a group in the following session.
Model simulations make clear the mechanisms that produce this advantage for the group-alone sequence. The cognitive facilitation that occurs in the group session carries over into the solitary session, where the brainstormer can continue without the hindrance of group inhibition. Whereas in most group sessions the potential cognitive advantages of the group interaction are overwhelmed by blocking and other inhibitory factors, an individual who follows a group session with a solitary session still has many of the ideas primed by the group active in short-term memory and is therefore able to use them to generate more ideas in a subsequent session. Furthermore, because a subsequent solitary session is not subject to blocking (and other inhibitory factors), the brainstormer can freely output those ideas. This facilitation shows up as a large “productivity spike” for solitary brainstormers in the second session in both the model simulations and the empirical data. This order effect is particularly strong when the initial group consists of heterogeneous members with differing knowledge of the task. The model simulations also predict that under ideal circumstances, a group session followed by a solitary session can even outproduce two back-to-back solitary sessions, although this advantage is expected to be small and has not yet been observed empirically. Simulations also indicate that a solitary brainstormer whose idea generation takes place after a group brainstorming session is likely to sample more categories from the task at hand than a similar brainstormer who works in two solitary sessions. This suggests an advantage of the group-alone sequence beyond any possible increases in overall productivity.
Another way to take advantage of group priming effects while reducing blocking effects is to have group members interact by writing and reading rather than speaking and listening. Although this may seem a sensible way to implement the group brainstorming formula, it does not seem to be a technique that is often attempted. Perhaps such “brainwriting” is a good example of a low-accessible (p.129) idea! We are so used to communicating orally when we are face to face that we just don't consider the alternatives. Paulus and Yang (2000) and Coskun (2000) tested interactive brainwriting as an alternative to interactive oral brainstorming. Instead of speaking their ideas as they occurred, group members wrote their ideas on a piece of paper and passed them on to the next group member, who read the idea on the slip of paper, added his or her own idea, and passed it on to the next group member. To prevent the slips of paper from becoming overcrowded and hard to read, brainwriters were asked to place slips that had accumulated four ideas in a pile in the center of the table. This assured that each brainwriter had the opportunity to look at each slip of paper at least once. This procedure resulted in interactive groups of brainwriters outproducing solitary brainstormers who wrote their ideas down on paper. These important results may be the first laboratory examples of face-to-face interactive groups outperforming an equal number of solitary brainstormers.
Although model simulations support the observation that interactive brain-writers can outperform an equal number of solitary brainwriters (Brown & Paulus, 2000), the simulation results are complex in some interesting ways. First, simulations predict that interactive brainwriting is not universally superior to individual brainwriting, but that it is most effective for heterogeneous groups (where members have differing knowledge of the brainstorming problem). More homogeneous groups actually show a productivity deficit in simulations. Second, performance of simulated brainwriting groups is an inverted U-shaped function of attention to the written ideas. Obviously, a brainwriter who does not read any of the ideas that are passed along will not benefit from the thoughts of his or her fellow brainwriters. A brainwriter who attends predominantly to the ideas of others will benefit from them to some extent, but not as greatly as someone who optimally balances the two goals of attending to the written ideas of others and following his or her own internal train of thought.
Our program of research has provided considerable support for our social information-processing model of group brainstorming. At the same time, many of our findings support recommendations made by Osborn (1957, 1963) many years ago. Osborn may have been a bit optimistic about the effectiveness of group brainstorming, but he was quite prescient about ways to optimize creativity in idea-sharing groups.
There are a number of pragmatic implications of our research findings and related theoretical framework. First of all, if groups are left to their own devices, are not trained, and interact in a nonstructured fashion, they will underperform in terms of creativity, innovation, and productivity. However, many tasks require group interaction or are most naturally done in face-to-face group situations. So it is important to develop some procedures to optimize the ability of groups to tap their creative potential. Our research suggests that we should pay (p.130) particular attention to procedures that affect the social influence and cognitive information processes in groups. To optimize motivation, there should be some degree of accountability and possibly some competition among group members and/or among groups. The development of group norms emphasizing high standards, open exchange of ideas, and attention to contrary points of view is important for the group's eventual success. One of the most difficult tasks in groups is to attend to the ideas being shared and simultaneously carefully access one's own knowledge base for relevant ideas. The information-exchange process should involve an alternation between individual and group sessions. After group sessions, individuals should immediately take some time to reflect on the ideas exchanged and note additional ideas that come to mind at that point or in subsequent days. These can be shared in subsequent meetings. Information-sharing sessions should include brief breaks to allow the processing of ideas and information exchanged and to encourage groups to take a new direction when they recommence. It is best to tackle one aspect of a problem at a time rather than focusing on an entire range of issues at once. All of these proposals imply that groups should have some degree of training to increase their effectiveness in group creativity. And again, most of these recommendations were originally considered by Osborn but have now been placed on a somewhat firmer empirical and theoretical foundation.
We finish by providing a list of recommendations for improving the effectiveness of brainstorming groups:
• Hold individuals accountable for generating their share of ideas.
• Set high goals for the group's performance.
• Minimize blocking by using writing or computer-based interaction.
• Alternate group and individual brainstorming sessions. Whenever possible, individual sessions should take place with minimal delay following group sessions.
• If possible, compose groups with members with complementary or heterogeneous sets of task knowledge.
• Motivate participants to attend carefully to each other's ideas.
• Include brief breaks in brainstorming sessions.
• Focus as much as possible on one aspect of a problem at a time.
• Be trained with an explicit set of rules about effective group interaction.
• Use trained facilitators or leaders to guide and motivate brainstorming groups.
When guidelines are used, group idea generation can be very effective, perhaps even more effective than individual idea generation.
We would like to thank Bernard Nijstad and Wolfgang Stroebe for their insightful comments on an earlier version of this chapter. Thanks are also due to Scott Burris (p.131) for helping to run many of the model simulations reported in this chapter while he was at the University of Richmond.
Amabile, T. M. (1996). Creativity in context. Boulder, CO: Westview Press.
Anderson, J. R. (1995). Cognitive psychology and its implications (4th ed.). New York: Worth.
Avolio, B. J., & Bass, B. M. (1988). Transformational leadership, charisma, and (p.132) beyond. In J. G. Hunt, B. R. Baliga, H. P. Dachler, C. A., Schriesheim (Eds.), Emerging leadership vistas (pp. 29–49). Lexington, MA: Lexington Books.
Avolio, B. J., & Bass, B. M. (1998). Individual consideration viewed at multiple levels of analysis: A multi-level framework for examining the diffusion of transformational leadership. In F. Danserau & F. Yammarino (Eds.), Leadership: The multilevel approaches (pp. 53–102). Stamford, CT: JAI Press.
Bandura, A. (2000). Exercise of human agency through collective efficacy. Current Directions in Psychological Science, 9, 75–78.
Baron, R. S. (1986). Distraction/conflict theory: Progress and problems. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 1–40). New York: Academic Press.
Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press.
Bass, B. M. (1998). Transformational leadership: Industry, military, and educational impact. Mahwah, NJ: Erlbaum.
Bass, B. M., & Avolio, B. J. (1994). Improving organizational effectiveness through transformational leadership. Thousand Oaks, CA: Sage.
Bennis, W., & Biederman, P. W. (1997). Organizing genius: The secrets of creative collaboration. Reading, MA: Addison Wesley.
Bonito, J. A., & Hollingshead, A. B. (1997). Participation in small groups. In B. R. Burleson & A. W. Kunkel (Eds.), Communication yearbook 20 (pp. 227–261). Thousand Oaks, CA: Sage.
Bousfield, W. A. (1953). The occurrence of clustering in the recall of randomly arranged associates. Journal of General Psychology, 49, 229–240.
Brown, V., & Paulus, P. (1996). A simple dynamic model of social factors in group brainstorming. Small Group Research, 27, 91–114.
Brown, V. R., & Paulus, P. B. (2000, November). Cognitive processes in group brainstorming. Poster presented at the 41st annual meeting of the Psychonomic Society, New Orleans.
Brown, V., Tumeo, M., Larey, T. S., & Paulus, P. B. (1998). Modeling cognitive interactions during group brainstorming. Small Group Research, 29, 495–526.
Burpitt, W. J., & Bigoness, W. J. (1997). Leadership and innovation among teams: The impact of empowerment. Small Group Research, 28 (3), 414–423.
Camacho, L. M., & Paulus, P. B. (1995). The role of social anxiousness in group brainstorming. Journal of Personality and Social Psychology, 68, 1071–1080.
Chemers, M. M. (2001). Leadership effectiveness: An integrative review. In M. A. Hogg & S. Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 376–399). Oxford: Blackwell.
Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407–428.
Coskun, H. (2000). The effects of outgroup comparison, social context, intrinsic motivation, and collective identity in brainstorming groups. Unpublished doctoral dissertation, University of Texas, Arlington.
Coskun, H., Paulus, P. B., Brown, V., & Sherwood, J. J. (2000). Cognitive stimulation and problem presentation in idea generation groups. Group Dynamics: Theory, Research, and Practice, 4, 307–329.
Cowan, N. (1995). Attention and memory: An integrated framework. New York: Oxford University Press.
(p.133) Diehl, M., & Stroebe, W. (1987). Productivity loss in brainstorming groups: Toward the solution of a riddle. Journal of Personality and Social Psychology, 53, 497–509.
Diehl, M., & Stroebe, W. (1991). Productivity loss in idea-generating groups: Tracking down the blocking effect. Journal of Personality and Social Psychology, 61, 392–403.
Dugosh, K. L., & Paulus, P. B. (2001). Cognitive and social comparison processes in brainstorming. Unpublished manuscript, University of Texas, Arlington.
Dugosh, K. L., Paulus, P. B., Roland, E. J., & Yang, H. C. (2000). Cognitive stimulation in brainstorming. Journal of Personality and Social Psychology, 79, 722–735.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117–140.
Fiedler, F. E. (1967). An integrative theory of leadership. Mahwah, NJ: Erlbaum.
Fiedler, F. E. (1978). Contingency and the leadership process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 11, pp. 167–225). New York: Academic Press.
Gammage, K. L., Carron, A. V., & Estabrooks, P. A. (2001). Team cohesion and individual productivity: The influence of the norm for productivity and the identifiability of individual effort. Small Group Research, 32, 3–18.
Geen, R. G. (1989). Alternative conceptions of social facilitation. In P. B. Paulus (Ed.), Psychology of group influence (pp. 15–51). Hillsdale, NJ: Erlbaum.
Gibbons, F. X., Blanton, H., Gerrard, M., Buunk, B., & Eggelston, T. (2000). Does social comparison make a difference? Optimism as a moderator of the relation between comparison level and academic performance. Personality and Social Psychology Bulletin, 26, 637–648.
Gibbons, F. X., & Gerrard, M. (1991). Downward comparison and coping. In J. M. Suls & T. A. Wills (Eds.), Social comparison: Theory and practice (pp. 317–345). Hillsdale, NJ: Erlbaum.
Goethals, G. R., & Darley, J. M. (1987). Social comparison theory: Self-evaluation and group life. In B. Mullen & G. R. Goethals (Eds.), Theories of group behavior (pp. 27–47). New York: Springer-Verlag.
Grossman, S. R., Rodgers, B. E., & Moore, B. R. (1989, December). Turn group input into stellar output. Working Woman, 36–38.
Guzzo, R. A., Yost, P. R., Campbell, R. J., & Shea, G. P. (1993). Potency in groups: Articulating a construct. British Journal of Social Psychology, 32, 87–106.
Horn, E. M. (1993). The influence of modality order and break period on a brainstorming task. Unpublished honors thesis. University of Texas, Arlington.
John-Steiner, V. (2000). Creative collaboration. New York: Oxford University Press.
Kerr, N. L., & Bruun, S. E. (1983). Dispensability of member effort and group motivation losses: Free-rider effects. Journal of Personality and Social Psychology, 44, 78–94.
Kramer, T. J., Fleming, G. P., & Mannis, S. M. (2001). Improving face-to-face brainstorming through modeling and facilitation. Small Group Research, 32, 533–557.
Leggett, K. L. (1997). The effectiveness of categorical priming in brainstorming. Unpublished master's thesis, University of Texas, Arlington.
Leggett, K. L., Putman, V. L., Roland, E. J., & Paulus, P. B. (1996, April). The effects of training on performance in group brainstorming. Paper presented at the Southwestern Psychological Association, Houston.
Lowe, K. B., Kroeck, K. G., & Sivasubramaniam, N. (1996). Effectiveness correlates (p.134) of transformation and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly, 7, 385–425.
Mandler, G. (1967). Organization and memory. In. K. W. Spence & J. T. Spence (Eds.), The psychology of learning and motivation (Vol. 1, pp. 327–372). New York: Academic Press.
Manz, C. C., & Sims, H. P. (1987). Leading workers to lead themselves: The external leadership of self-managing work teams. Administrative Science Quarterly, 32, 106–129.
Meyer, D. E., & Schvaneveldt, R. W. (1971). Facilitation in recognizing pairs of words: Evidence of a dependence between retrieval operations. Journal of Experimental Psychology, 90, 227–234.
Mitchell, K. A. C. (1998). The effect of break task on performance during a second session of brainstorming. Unpublished master's thesis, University of Texas, Arlington.
Mullen, B., Johnson, C., & Salas, E. (1991). Productivity loss in brainstorming groups: A meta-analytic integration. Basic and Applied Social Psychology, 12, 3–23.
Neely, J. H. (1991). Semantic priming effects in visual word recognition: A selective review of current findings and theories. In D. Besner & G. W. Humphreys (Eds.), Basic processes in reading: Visual word recognition (pp. 264–350). Hillsdale, NJ: Erlbaum.
Nosofsky, R. (1984). Choice, similarity, and the context theory of classification. Journal of Experimental Psychology: Learning, Memory, & Cognition, 10, 104–114.
Offner, A. K., Kramer, T. J., & Winter, J. P. (1996). The effects of facilitation, recording, and pauses on group brainstorming. Small Group Research, 27, 283–298.
Osborn, A. F. (1957). Applied imagination. New York: Scribner.
Osborn, A. F. (1963). Applied imagination (2nd ed.). New York: Scribner.
Oxley, N. L., Dzindolet, M. T., & Paulus, P. B. (1996). The effects of facilitators on the performance of brainstorming groups. Journal of Social Behavior and Personality, 11, 633–646.
Parnes, S. J., & Meadow, A. (1959). Effect of “brainstorming” instructions on creative problem-solving by trained and untrained subjects. Journal of Educational Psychology, 50, 171–176.
Paulus, P. B. (1983). Group influence on individual task performance. In P. B. Paulus (Ed.), Basic group processes (pp. 97–120). New York: Springer-Verlag.
Paulus, P. B. (2000). Groups, teams and creativity: The creative potential of idea generating groups. Applied Psychology: An International Review, 49, 237–262.
Paulus, P. B., Brown, V., & Ortega, A. H. (1999). Group creativity. In R. E. Purser & A. Montuori (Eds.), Social creativity in organizations (pp. 151–176). Cresskill, NJ: Hampton.
Paulus, P. B., Dugosh, K. L., Dzindolet, M. T., Coskun, H., & Putman, V. L. (2002). Social and cognitive influences in group brainstorming: Predicting production gains and losses. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 12, pp. 300–325). London: Wiley.
Paulus, P. B., & Dzindolet, M. T. (1993). Social influence processes in group brainstorming. Journal of Personality and Social Psychology, 64, 575–586.
Paulus, P. B., Dzindolet, M. T., Poletes, G., & Camacho, L. M. (1993). Perception of performance in group brainstorming: The illusion of group productivity. Personality and Social Psychology Bulletin, 19, 78–89.
Paulus, P. B., Larey, T. S., & Dzindolet, M. T. (2000). Creativity in groups and teams. (p.135) In M. Turner (Ed.), Groups at work: Advances in theory and research (pp. 319–338). Hillsdale, NJ: Erlbaum.
Paulus, P. B., Larey, T. S., Putman, V. L., Leggett, K. L., & Roland, E. J. (1996). Social influence process in computer brainstorming. Basic and Applied Social Psychology, 18, 3–14.
Paulus, P. B., & Yang, H. C. (2000). Idea generation in groups: A basis for creativity in organizations. Organizational Behavior and Human Decision Processes, 82, 76–87.
Pinsonneault, A., Barki, H., Gallupe, R. B., & Hoppen, N. (1999). Electronic brainstorming: The illusion of productivity. Information Systems Research, 10, 110–133.
Prince, G. M. (1970). The practice of creativity. New York: Harper and Row.
Putman, V. L. (1998). Effects of facilitator training and extended rules on group brainstorming. Unpublished master's thesis, University of Texas, Arlington.
Putman, V. L. (2001). Effects of additional rules and dominance on brainstorming and decision making. Unpublished doctoral dissertation, University of Texas, Arlington.
Raaijmakers, J., & Shiffrin, R. (1981). Search of associative memory. Psychological Review, 85, 59–108.
Rawlinson, J. G. (1981). Creative thinking and brainstorming. New York: Wiley.
Roy, M. C., Gauvin, S., & Limayem, M. (1996). Electronic group brainstorming: The role of feedback on productivity. Small Group Research, 27, 215–247.
Sanders, G. S., Baron, R. S., & Moore, D. L. (1978). Distraction and social comparison as mediators of social facilitation effects. Journal of Experimental Social Psychology, 14, 291–303.
Shepherd, M. M., Briggs, R. O., Reinig, B. A., Yen, J., & Nunamaker, J. F., Jr. (1995–1996). Invoking social comparison to improve electronic brainstorming: Beyond anonymity. Journal of Management Information Systems, 12, 155–170.
Shepperd, J. A. (1993). Productivity loss in performance groups: A motivation analysis. Psychological Bulletin, 113, 67–81.
Sherwood, J. J. (1998). Modeling cognitive facilitation in brainstorming groups. Unpublished master's thesis, University of Texas, Arlington.
Smith, S. M. (1995). Fixation, incubation, and insight in memory and creative thinking. In S. M. Smith, T. B. Ward, & R. A. Finke (Eds.), The creative cognition approach (pp. 135–156). Cambridge, MA: MIT Press.
Sosik, J. J., Avolio, B. J., & Kahai, S. S. (1997). Effects of leadership style and anonymity on group potency and effectiveness in a group decision support system environment. Journal of Applied Psychology, 82, 89–103.
Sosik, J. J., Avolio, B. J., & Kahai, S. S. (1998). Inspiring group creativity: Comparing anonymous and identified electronic brainstorming. Small Group Research, 29, 3–31.
Sosik, J. J., Kahai, S. S., & Avolio, B. J. (1998). Transformational leadership and dimensions of creativity: Motivating idea generation in computer-mediated groups. Creativity Research Journal, 11, 111–121.
Stanne, M. B., Johnson, D. W., & Johnson, R. T. (1999). Does competition enhance or inhibit motor performance: A meta-analysis. Psychological Bulletin, 125, 133–154.
Steiner, I. D. (1972). Group process and productivity. San Diego: Academic Press.
Sutton, R. I., & Hargadon, A. (1996). Brainstorming groups in context. Administrative Science Quarterly, 41, 685–718.
(p.136) Taylor, D. W., Berry, P. C., & Block, C. H. (1958). Does group participation when using brainstorming facilitate or inhibit creative thinking? Administrative Science Quarterly, 3, 23–47.
Tulving, E., & Pearlstone, Z. (1966). Availability versus accessibility of information in memory for words. Journal of Verbal Learning and Verbal Behavior, 5, 381–391.
Turner, J. C. (1978). Social categorization and social discrimination in the minimal group paradigm. In H. Tajfel (Ed.), Differentiation between social groups (pp. 101–140). London: Academic Press.
Williams, K. D., & Karau, S. J. (1991). Social loafing and social compensation: The effects of expectations of coworker performance. Journal of Personality and Social Psychology, 61, 570–581.
Zajonc, R. B. (1980). Compresence. In P. B. Paulus (Ed.), Psychology of group influence (pp. 35–60). Hillsdale, NJ: Erlbaum.
(1.) Nijstad, Diehl, and Stroebe (this volume) also present a model of cognitive processing in group brainstorming. The model SIAM (Search for Ideas in Associative Memory) is built on Raaijmakers and Shiffrin's (1981) SAM model (Search of Associative Memory) and therefore, like the model presented in this chapter, is also based on an associative memory framework. Thus, there are a number of similarities in the two models, with perhaps the major distinction being one of emphasis: our associative memory model emphasizes the structural properties of semantic networks, and SIAM emphasizes the active search process that takes place when a brainstormer (either alone or in a group) is seeking to generate novel ideas. The emphasis on an active search process in SIAM may make it a better candidate for exploring the process of constructing novel combinations of existing ideas.
(2.) This matrix representation of long-term semantic memory has some obvious similarities to existing mathematical memory models, such as Raaijmakers and Shiffrin's (1981) SAM model. SAM, in particular, represents individual ideas as “images” that are connected to each other with varying strengths; these strengths in turn determine the probability that a given image will activate another, related image. So, like the standard semantic network representation, SAM represents memory at the level of individual concepts. If a category is viewed as a collection of individual ideas, our matrix model can be thought of as a higher-level representation of SAM-like images: the entries in the matrix can presumably be derived from the association strengths among individual images in SAM (cf. Nosofsky's 1984 “mapping hypothesis”).
(3.) Like SAM, which can capture the temporal nature of memory recall (e.g., serial position effects such as primacy and recency), our associative memory model captures some of the dynamic nature of individual and group brainstorming. The decrease in idea generation that takes place as a brainstorming session progresses is built into the model by decrementing the probability of generating another idea from a category each time a new idea is generated from that category. At the start of a session, the probability of staying in a category is higher than the probability of switching to any other category, so ideas from within the same category will tend to follow each other. Also, because a decrease in the probability of staying within a category is accompanied by an increase in the probabilities in the null category (which represent the likelihood of generating no new idea), switching to a new category will, on average, take longer than coming up with an idea from the same category. The result is that ideas tend to come in clumps, with more “blank time” between ideas from different categories than between ideas from the same categories. The SIAM model described in Nijstad et al. (this volume) accounts for these temporal patterns as well.