I have finished reading “Wiser”, the latest book by the North American jurist and academic Cass Sunstein, co-authored by the Chicago University professor Reid Hastie. It was published in January 2015, so the print is still quite fresh. The book is mainly of interest because it covers factors that give rise to (and can inhibit) Groupthink.
As you may remember, “Groupthink” is a term coined in the seventies by the psychologist Irving Janis, naming those situations where individuals participating in a group adapt and submit to the collective opinion even if it differs from their own point of view. The more cohesive the group the stronger the bias, because the social (and informational) pressure that generate cohesion affect the individuals’ capacity to make good use of their private information sources, thus gravitating to the groups’ central opinion. The consequences of this behavior are negative. Groups end up making bad or irrational decisions because the diversity of opinions of the individual group members are not aggregated efficiently.
“Wiser” addresses this issue in two parts. The first half of the book analyses the factors that lead to different cognitive biases when groups are at work as a collective. The second presents different palliative measures for the Groupthink effect.
This subject has been approached by many authors. James Surowiecki, in “The Wisdom of Crowds” analyses this phenomenon in some depth (with plenty of examples), reminding us once more that “as a group becomes more cohesive, the individual becomes more dependent“. Reducing the adverse effect of Groupthink is one the greatest challenges in the practice of Collective Intelligence.
There are a variety of causes that give rise to Groupthink, and I’ll go through many in other posts on this Blog. Today I want to concentrate on one of the lesser known because it is a good starting point to shine some light on the complexity of the problem. It occurs when the group only recognizes and prioritizes information that is shared and available for all group members. This is what experts call the “Common Knowledge Effect”, where a group tends to reject or ignore information or data that is only in the hands of a few members with novel viewpoints. To be more graphic, it is as if statistics determine an idea’s reputation: “if everybody has this bit of information than it is more valuable than information in the hands of just a few”, in other words, “we don’t know what you know, therefore it’s worthless”.
The book provides several examples of scientific research that prove that information shared with everyone is much more influential and has a larger impact on the group’s decision than information shared by just a few; so we can derive our first practical lesson: it is wise and convenient to make sure that all relevant information is distributed to the whole group before tackling a problem.
Sunstein explains that there are group members that are “cognitively central”, the knowledge that they handle is usually shared with the rest of the group. But there are also individuals that are “cognitively peripheral”, that is, they access information that others don’t. Knowing this, intelligent groups make good use of the latter because the unique, different, knowledge that they can bring to the table is precisely what the group needs to hear most.
Points of view that are discarded or ignored are technically known as “hidden profiles”. So Collective Intelligence is challenged to lower the number of these profiles. This will not happen spontaneously and we can’t trust improvisation to make it possible. As I like to say, it is a matter of designing the participatory architecture. The group must activate mechanisms that amplify “cognitively peripheral” profiles, especially (and this is key) if these individuals are “low status”, because in this case the probability of them losing their voice increases significantly.
Group members that handle unique or unusual information and are, at the same time, low status (with little authority because of lack of prestige or hierarchy or experience within the group) take on more reputational risk when conveying their knowledge. In short “they are mavericks”, they are pointed out. So in many cases they keep quiet relinquishing the defense of their point of view if there are no mechanisms to encourage them to speak out.
Leadership should therefore double check the following: 1) Distribute all relevant information to all of the group members, 2) Pay special attention (explicitly) to “peripheral” individuals (with unusual information or logic) because they are the ones that know what the rest doesn’t.
Notes: This post was translated into English with the help of Peter Hodgson. The image of the post belongs to the album of Daryl Marquardt in Flickr. Read this post in Spanish (Lee este post en Español)