Evaluating Behavior Priming Research: Three Observations and a Recommendation

We discuss three observations about the state of the field of behavior priming. The first is that there are many more empirical demonstrations of behavior priming than may be apparant at first sight. The second is that some people doubt the validity of behavior priming effects because they are “counterintuitive,” and we argue that this reasoning is subjective, often circular, and sometimes based on an underappreciation of relevant theories. The third is that we concede that—just as in many other areas—publication bias and the use of researchers' degrees of freedom have presumably led to a somewhat distorted literature because they caused what we call “information leakage.” Because of some of the habits in response to publication bias, our field knows much less about moderators and boundary conditions than we could have known. We conclude that although replication efforts can be very useful, the only true solution to further improve our field is to stop practices, such as the liberal use of researchers' degrees of freedom, that scientists have adopted in response to the pressure to publish only statistically significant results.