This study stems from research that has shown many post-conviction DNA exonerations have involved a failure of investigators to take seriously the suspects' alibis that would have proven their innocence. The current study's hypothesis was confirmed; study participants who had to provide their own alibis in a particular circumstance rated a criminal suspect's alibi as more believable, compared to study participants who evaluated the suspect's alibi without being required to provide an alibi for themselves. This finding is consistent with earlier research, which shows that people reference their own knowledge and emotional states in determining what others know or feel. This suggests that if detectives extrapolate from their own experience (i.e., that their own alibis are weak or difficult to prove), then they are more likely to recognize that weak alibis can stem from oversights; memory inaccuracies; and the failure to give attention to times, dates, behaviors, and persons involved in activities in which one participated at an earlier time. Study participants (N=147) were randomly assigned to either evaluate a suspect's alibi before generating their own alibi (evaluate-first) or generate their own alibi before evaluating the suspect's alibi (generate-first). Participants provided alibis from either 3 days previous or 30 days previous. In Experiment 2, participants (N=255) were randomly assigned to either generate-first, evaluate-first, or read-experience; the latter group read about alibi-generation difficulty. Half of the participants were primed to think empathetically, and half were not. All participants evaluated the believability of the suspect's alibi as well as their own alibis. 3 tables and 25 references