ยป Homework Review: Why Societies Need Dissent
Cass Sunstein, the legal scholar, wrote this before he teamed up with Richard Thaler to write Nudge. I had to read the first half of it for class. It obviously worked as a draft for some of the ideas that went into Nudge, but the first half is not worth reading itself. I get the feeling that the book is a paper that’s been padded out to book length.
Sunstein’s core idea is that in group decision making, people’s private information is supressed. There are two possible reasons:
- The decision-making process is not deliberative. For example, when you get a new doctor and they read your chart, it shows previous doctor’s decisions but not their reasoning. There’s no way to tell if a previous doctor recommended a treatment with reservations or confidently.
- People have goals besides optimal group decisions. The most problematic is the goal of social cohesion. When a group is making a decision, members will withhold private information if they believe the benefit of revealing that information is lower than the social cost of dissenting.
Rather than real-world examples, most of the evidence given for these issues is from a single set of economics experiments:
The experimenter presents an urn that has three rocks, either two white and one black or the opposite. Each member of the group secretly examines one rock and then announces to the group a guess of what rocks are in the urn. For example: if the first examiner guesses that the urn contains two white rocks and the second examiner sees a black rock, then the second examiner knows that there’s an even chance that it’s either kind of urn, but they must guess two black rocks to pass the right information to the third examiner. It’s like bridge.
Sunstein’s conclusion from all this is that dissent is important to improve decisions but also prevents consensus. I didn’t get to the part where he recommends what to do about it, if there is such a part. For example, I can imagine a decision making process that starts out anonymous to encourage dissent then moves to face-to-face to encourage consensus.



I don’t agree with this at all!
Seriously though: that experiment sounds like it assumes humans are rational actors.
FAIL!
Jack
21 May 10 at 12:41 am
Well, the purpose of the experiment is to show that human bounded-rationality is skewed in a particular way. Basically, that they have a heuristic that says to trust the decisions made by other people. But I agree with you: how many people understand probability well enough that they’d get it right with perfect information?
Jared
21 May 10 at 10:57 am