A Defense of Groupthink?

According to a recent study (seen via BoingBoing), one of the best incentive structures is punishing disagreement. Can we punish these scientists for disagreeing with the theory that groupthink and censorship are bad?

On the other hand, after I looked at the paper, I noticed the following speculation about the incentive structure they call Bayesian Truth Serum (BTS), another way to reward agreement:

When compared with the other treatments and control conditions, BTS likely had two effects: (a) it created some confusion among subjects about how exactly they were being evaluated; and (b) it created an incentive for subjects to think carefully about the respones of other subjects. The combination of confusion and cognitive demand probably elicited greater engagement with the question, and this engagement in turn probably drove better performance.

In other words, these incentive structures might have avoided the common idiocy “I don't have to think; I have the right to my thoughtless opinion.” I suppose groupthink might be an improvement over nothink.

Addendum: After some more thinking on this, I realized that BTS will reward agreement but it will punish attempted agreement. If you try picking the most popular answer, you are more likely to lose.