Some AI research led me to an interesting article about a "Bayesian truth serum." I won't bore you with Bayes Theorem, but follow the link if you want to read about it. What the news articles has to say about it is that sometimes questions can be constructed in pairs that help shed more information on the answers.
For example, he describes a situation where two paintings are viewed by a group of 10 people who are then asked, privately, to pick their favourite. Seven people say they prefer painting A, while three vote for painting B. If, on the second question, all 10 people said they thought everyone else would prefer painting A, then those three people expressing a personal preference for painting B might be thought of as a safer bet for having told the truth. That is because, argues Prelec, despite what they thought was more popular, those individuals still chose the other painting.
Personality tests have used a related technique for quite some time, to tell if someone is trying to game the test. This approach could potentially be quite different, and could have good applications to market research or job interviews. I can't really think of any good question pairs off the top of my head that would help in those areas, but if you can, please leave them in the comments.