Edge.org's annual question for 2008 is: "What have you changed your mind about? Why?" There are a number of extremely interesting answers; I'm going to highlight only a small selection.
Susan Blackmore's de-conversion from parapsychology and the paranormal is well known, but it's still interesting reading. The tale is highly recommended reading to those not yet in the know.
David Buss says he now thinks female sexual psychology is several orders of magnitude more complex than he previously thought.
Michael Shermer is no longer a behaviorist, he has "thus changed [his] mind about this theory of human nature in its extreme form. Human nature is more evolutionarily determined, more cognitively irrational, and more morally complex than I thought." Welcome to the fold, Michael.
Marc Hauser says he has "lost the faith... in the power of the adaptive program to explain or predict particular design features of human thought." These, according to Hauser, include "language, morality, music, and mathematics".
David G. Myers has changed his mind about, among other things, whether children are blank slates, repressed memory syndrome, electroconvulsive therapy and whether birth-order affects personality.
Scott Atran once "thought that individual cognition and personality, influences from broad socio-economic factors, and degree of devotion to religious or political ideology were determinant." Now, however, he "see[s] friendship and others aspects of small group dynamics, especially acting together, trumping most everything else." That is, he now thinks "fictive kin" plays a key role in religious politics.
Robert Trivers has come to believe that understanding human self-deception requires a surprisingly deep understanding of biology.
Judith Rich Harris argues generalization - e.g. fear of a white rat generalizing to a fear of white furry things - is the exception rather than the rule.
Steven Pinker has developed doubts about evolutionary psychology's assumption that human evolution ground to a halt at the beginning of the Holocene. Evidence has accumulated that large parts of the human genome has been under strong recent selection pressure and the result, speculates Pinker, "could be evolutionary psychology on steroids. Humans might have evolutionary adaptations not just to the conditions that prevailed for hundreds of thousands of years, but also to some of the conditions that have prevailed only for millennia or even centuries."
Helena Cronin has changed her mind about how best to account for sex differences.
Geoffrey Miller now thinks the intellectual division of labor has gone too far: that behavioral scientists should start discussing human nature with ordinary people to help them come up with testable hypotheses. Says Miller: "Marine drill sergeants know a lot about aggression and dominance. Master chess players know a lot about if-then reasoning. Prostitutes know a lot about male sexual psychology. School teachers know a lot about child development. Trial lawyers know a lot about social influence." So "whenever we try to understand human nature in some domain, we should identify several groups of people who are likely to know a lot about that domain already, from personal, practical, or professional experience."
Dan Sperber discusses how reading one of Leda Cosmides' papers convinced him to become an evolutionary psychologist.