The goal of this study is to present and validate a simple method for accounting for peer selection on offending based on a respondent's self-reported preferences for friends who engage in criminal behavior. Using primary panel data (n = 611), having a preference for peers who offend (the measure of peer selection) relates positively and significantly to offending behavior. The selection measure, which carries the advantage of being closely aligned to criminological theory, renders the peer offending/perso nal offending relationship nonsignificant. Our selection variables also out perform a more traditional means of capturing peer selection effects.