Choice-reaction time (CRT) is an experimental information-processing paradigm. Based on an interference effect in visual attention, the CRT method has been shown to be suitable for measuring sexual orientation in men and women. The present study assessed the potential of the CRT to identify deviant (i.e., pedophilic) sexual interest. Participants were patients from forensic-psychiatric hospitals: 21 child molesters and 21 non-sex offenders. The dependent variable was reaction time in an ostensible seek-and-locate task (i.e., identifying the position of a dot superimposed on a picture of a person). There was an interaction effect between stimulus age category and participant group status: Child molesters took longer to respond to pictures of children relative to pictures of adults. Non-sex offenders showed an opposite pattern (i.e., longer reaction times with pictures of adults than with pictures of children). In addition, the data supported the notion of sexual content induced delay: Subjects took longer for the task with nude stimuli than with clothed ones. A subtractive preference index, derived from the reaction times for child and adult stimulus material, allowed distinguishing participants from both groups almost perfectly (ROC-AUC = .998). We conclude that a match of sexual interest with properties of visual stimuli led to a cognitive interference effect: Attentional resources were drawn from the ostensible task of locating the dot towards exploring the picture. This opens up the possibility of using this interference effect (i.e., the delay of response times) for diagnostic purposes.