Bingo. A study I can't find after trying on Google took people with objectively wrong opinions and payed them for accuracy. Magically, people corrected their own opinions when they had a financial incentive to do so.
Interesting. I'd like to read that study, if you ever figure out which one it was.
Meanwhile, I wonder if this was a case of people figuring out what the testers wanted to hear, rather than actually altering the beliefs they held.
Works for a while, not forever.