Relative viewing times (VTs)—time required to view and evaluate sexually salient images—discriminate individuals with a sexual interest in children, as indirectly indexed by their history of sexual offending against children, from those without such history. In an online sample of 652 fathers, we measured VTs and sexual attraction ratings to child and adult images. We assessed participants’ sexual offending history and propensity (self-reported likelihood to have a sexual contact with a child, a non-consensual sexual contact with an adult, and propensity toward father–daughter incest). In contrast with VT studies involving clinical or forensic samples, VTs and attraction ratings failed to discriminate participants with a sexual offending history. VTs successfully distinguished participants with a propensity to sexually offend against children but failed to identify those with a propensity toward incest. Conversely, attraction ratings distinguished participants with a propensity toward incest but failed to identify those with a propensity to sexually offend against children. Correlations between VTs and attraction ratings were small. Results illustrate, for the first time, the distribution of VT measures in community fathers, support the feasibility of online administration of VT tasks to detect propensity to sexually offend against children, and indicate that sexual interest in children and incest propensity are distinct.