Freakonomics Steven Levitt is not a fan of priming, and failed to produce satisfying results in a study of his own. But he may be becoming convinced…

• ‘Priming’ is the idea that you can give a person hints and stimuli to influence their behaviour in the short term. Ask someone if they are a woman before they do a math test, and the cultural perceptions of women being worse at math will influence their performance.
• Levitt tried his own priming study, that failed to produce results.
• However, a study has been done going into prisons and asking prisoners either what their crime was, or how much t.v. they watch. Prisoners then flipped coins in secret, and received payment based on the number of heads they recorded.
• The results? Control non-prisoner populations lied 6% of the time. Prisoners asked about t.v. lied 10%, and prisoners asked about their crime lied 16% .

The article over here is pretty short, but contains a fuller methodology, links to the Freakonomics study and the priming prison study.

Source: Freakonomics