Wednesday, 18 May 2011

The much maligned group brainstorm can aid the process of combining ideas

Research on group creativity shows consistently that the same people come up with more ideas working on their own than they do when brainstorming together. But perhaps it's time to move beyond this striking yet superficial discovery. After all, having a list of initial ideas is not the end of the creative process. A new study by Nicholas Kohn and colleagues has focused on the creative task of idea combination, finding that in this context groups do have advantages over individuals working alone.

One hundred and eight student participants formed groups of three working at computer terminals located apart (this set-up was used to rule out the influence of various social factors that emerge in face-to-face situations). The participants' ten-minute task was to come up with fresh ideas for how to improve their university. Some of the groups of three shared their ideas electronically - that is, each individual could see the ideas of their two fellow team members appear on their own screen as they worked. Other groups worked alone, each individual entirely cut off from their two team members.

The next stage was about idea combination. All participants, whether they previously worked alone or not, now had access to a list of their own existing ideas and the already proposed ideas of their team members. For the next fifteen minutes participants attempted to combine these existing ideas into novel concepts, or to combine an existing idea with a new one. Crucially, half the participants (whether they previously worked alone or collaboratively) now did the combining on their own; the other half could see their team members' newly combined ideas appear on-screen as they worked.

Consistent with past research, participants who worked alone in the first phase came up with more ideas than those who worked cooperatively with their team members. However, team working was more successful in the second, idea combination phase. Although participants working on their own came up with more combined ideas, it was the combined ideas produced by participants working together that were rated by independent judges as being more useful.

Another finding was that participants who worked alone in the first phase were more likely to use other people's ideas to form novel combinations in the second phase (rather than just combining their own earlier ideas), perhaps because they were seeing them for the first time and therefore finding them more stimulating.

A second study was similar to the first except the participants were asked to form newly combined ideas out of existing ideas from an external source (ie. not generated by themselves or their team members). The topic was as before - how to improve the university. Some groups worked with ideas categorised as common, others with rare ideas. This time the collaborators sat around a table and followed a "brain writing" technique - each time they conceived of a new idea combination they wrote it down on a piece of paper and passed it to their neighbour, who rated its usefulness. The purpose of this was to make sure collaborating participants engaged with each others' ideas.

Again, individuals working alone generated more freshly combined ideas than individuals working collaboratively - this was unsurprising since the brain writing process is time consuming. However, participants working collaboratively with rare material came up with combined ideas that judges rated as more novel and feasible, than did participants working alone. And collaborating participants working with common material came up with combined ideas rated as having more impact. This result shows again that there are times in the creative process when working collaboratively has advantages.

'Our results provide a fertile basis for future studies to examine the factors that influence this process and enhance the ability of groups to generate combinations that are both original and useful,' Kohn and his team concluded.

ResearchBlogging.orgKohn, N., Paulus, P., and Choi, Y. (2011). Building on the ideas of others: An examination of the idea combination process. Journal of Experimental Social Psychology DOI: 10.1016/j.jesp.2011.01.004

This post was written by Christian Jarrett for the BPS Research Digest.


  1. I am pleased to see that Paulus and his colleagues have finally come up with an advantage of brainstorming after bashing it all these years. This research always bewilders me -- in part because my colleagues and I have studied brainstorming in context, not the lab -- as when you see how it is used in a real company by people who do real creative work and have real skills, the focus on the number of ideas generated per-person on a given period of time (note this study used a different dependent variable, quality of the ideas, not quantity)turns out to be trivial -- lab researchers like it because it is easy to measure and is sensitive to little tweaks made in the lab. But things like using brainstorming sessions to refresh and spread solutions around the company (we call this refreshing the organizational memory -- going to brainstorms is especially crucial for newcomers), creating a constructive status contest where people compete to generate the best ideas (important to cultures of innovation), giving people interesting and emotionally satisfying work (the finding that people enjoy brainstorming with others more than alone is dismissed as trivial by brainstorming researchers like Paulus, but in the real world here in Silicon Valley, giving people work they like to do is important for recruiting, retaining and energizing them), and finally, as gaining commitment for clients and senior executives is important, the experience of hearing a group generate ideas impresses these important visitors far more than listening to one smart person after another brainstorm. And,oh, by the way, it is mighty hard to combine your ideas with others ideas when your brainstorm alone because there are no other people there! So while I applaud a lab study that shows, finally, that there is some advantage to brainstorming alone, if you look at research done in real organizations where people do brainstorming with real consequences and have real bosses and real problem, although it is too messy for people in the lab to deal with in any meaningful way, and the main measure -- quantity -- is imply silly.

    P.S. Although this study seems to model what happens in real places that do brainstorming than other experiments, once again, it uses subjects who have no significant past experience doing brainstorming alone or in groups pr -- very crucially -- leading brainstorming sessions. These are both skills that take time to learn to do well and -- even at the most creative companies -- there are big differences in skill. This is another reason that the assertion that lots of studies show that brainstorming doesn't work are rather silly because whatever Paulus and others study in experiments has little resemblance to what is actually done in real companies.

    Bob Sutton
    Stanford University

    The main study I am talking about is:

  2. Thanks Bob, it's really interesting to hear a different perspective from the 'real world' of business.

  3. It is also the real world of evidence and theory. My take on brainstroming researchers, especially Paul Paulus and his colleagues-- is they have constructed a fantasy about what brainstorming is and how it is used that is detached from reality in many ways, and therefore useless to understanding the real problem, but useful to them on that it produces results that can be published in academic journals. I would add that this opinion of such research is held be some rather famous experimental psychologists. When I presented my ethnographic research at the Stanford psychology department perhaps a decade ago renowned faculty including the late Robert Zajonc held a similar, and in fact, more negative opinion of how these researchers had framed the problem. I am a huge fan if experiments, and accepted many as a journal reviewer and editor, but the way these folks frame and use these data are misguided.

  4. Strange Medicine3:34 pm

    Wonder if this is contextual - doesn't seem to be particularly beneficial in the case of jury decision-making. A recent, as yet unpublished, study of ours demonstrates that the same biases that individuals have in thinking through a case show up when a jury are thinking through a case. Yes, they are mock juries, but evidence from conviction rates suggest there may be similar processes going on plus jury members are not expert jury members, so are less likely to be as different as is the case between the student sample in the described study and Bob Sutton's description of expert business brain stormers.

  5. Nicholas Kohn2:28 pm

    Where Bob & I disgree on is this use of emperical data obtained from laboratory experiments. He advocates for the use of field research in business settings. I have for the most part focused my efforts in lab experiments. It's my belief that both lab and field experiments are necessary to provide a complete picture.

    Field study provide external validity but are plagued by confounding variables and the difficulty in studying some mechanisms. Lab experiments potentially lack external validity, but allow me to study the cognitive and social mechanisms that underlie brainstorming.

    It is not Paulus's nor my intention to "bash brainstorming." Rather, it has been our focus to study the cognitive and social mechanisms that can lead to deficits in brainstorming or facilitate brainstorming. By understanding these, we can then offer recommendations on how to enhance brainstorming.

    Nicholas Kohn

  6. I am not opposed to lab studies, and I appreciate their strengths. If you read my academic and popular writings, I use them all the time and have supervised dissertations that use them. Rather, my objection is arises when people generalize beyond the data. In the mid 1990's, I presented an early version my brainstorming research to a room that contained some mighty famous experimental psychologists including the late Bob Zajonc, Mark Lepper, and Lee Ross -- they thought the premise of the brainstorming experiments were absurd and Ross in particular argued that it was a perfect example of using the lab to study something that is responsive to manipulation but meaningless. I actually think yours is one of the rare studies that gets closer to what matters in the real world. We agree completely that both the lab and field are needed in harmony, but both lab and field researchers need to be careful in their claims. I don't believe that lab research supports he claim that "brainstorming doesn't work even though I have heard many psychologists say it and it is parroted in a recent New Yorker, just as I believe that my work does not allow me to say it does work -- only that more rigorous research needs to frame it differently, as your study does.

    Bob Sutton


Note: only a member of this blog may post a comment.