last updated 11/11/2014

The Surprising Failures of Deliberative Groups
Sunstein' Infotopia.  pp. 45-73

Thesis/ Themes

• While we, as citizens, assume decision-making by deliberating groups (i.e. groups that meet and discuss face-to-face) is good and effective, such groups can go very wrong.

  • While majority opinion tends to rule in deliberative settings, the "truth" influences decisions - but only somewhat

  • Groups sometimes make good decisions (esp. if the majority starts out favoring the good decision) but often they don't.

  • Groups can end up making worse decisions than any individual would have made.  Consider situations as diverse as the Bay of Pigs fiasco, student group hazing incidents, the GM ignition switch recall.

  • Groups work best at getting the "right" answer if the right answer is obvious when you see it (due to the effects of the Condorcet Jury Theorem).

  • Deliberating groups vary a lot in effectiveness (see above points); thus, group deliberation is not always a good idea.

  • Sunstein assumes (mostly correctly) that an essential to making good decisions is being careful to pool the best information available.

Key Concepts

• In groups where there is substantial initial agreement (e.g. mostly liberals or conservatives), views become more extreme as a result of discussion.  (this is called "risky shift.")

• Groupthink, based on the research of Irving Janis into the causes of the Cuban Bay of Pigs Invasion fiasco, illustrates how, in highly cohesive grooups, social pressure and witholding information can cause disasterous mistakes.

• Sunstein tends to presume that the principle function of deliberation is to combine information and expand the range of thinking.  His take on the U.S. Constitution is that pooling information is what the structures of government are intended to do.  But there may be more to it.  (What?)

• A highly confident group member may lead the group in his/her direction, for better or worse (e.g. 12 Angry Men).

• Sunstein suggests small group outcomes tell us about larger groups, but is that true?

• Herb Simon's concept of "satisficing" [The tendency of individuals and groups to gain just enough information to develop a preference for a decision, often ends the search for information and leads to efforts to avoid contradictory information].  This is at least partially a feature of the human desire to reduce ambiguity - whether by individuals or in groups.  This problem is amplified by the desire to be "effcient."

• SHARED knowleged plays a much bigger role in decision-making than UNSHARED knowledge even when the unshared knowledge is better.  Groups are less likely to ignore unshared knowledge when they believe there is a demonstrably correct answer to the problem they are facing.

• The two most important reasons for deliberative failure  (Low status members are most affected by these.)

  • Informational influences -- people defer to others (and don't share their information) when they have the impression others are right/more confident.  (informational signals)

  • Social influences -- people defer to the group/others when they fear their statements will be punished, ridiculed or disliked.
     

  • Groups function better and share information better when the benefits of good decision-making result in meaningful, real rewards for the individual members.