The issue of how a problem behavior (the offense chain) and relapse (reoccurrence of a problem behavior following a period of abstinence) can best be distinguished may be resolved by asking whether there are sound clinical, theoretical, or empirical reasons for making this distinction. Such an analysis shows that the sets of factors that underlie etiology and relapse have similarities. Such similarities suggest that it may be more profitable to focus more directly on the problem behavior process itself. New pathways can always be added to the model. The suggested offense model involves seven stages. Stage 1 includes proximal background factors, such as the offender's perception of his general circumstances. Stage 2, distal planning, involves three possibilities: covert planning, chance contact, and explicit planning, which may be associated with positive affect. Stage 3 involves nonsexual contact with the victim for the purpose of offending, and stage 4 involves the offender re- evaluating his current circumstances and often changing his perception of his relationship with the victim in order to allow offending to occur. Stage 5, proximal planning, concerns the immediate precursors to the sexual offense and involves behaviors such as getting into bed with the victim. Stage 6 is the sexual offense, and stage 7 is the assessment process that occurs in reaction to having committed the offense. In stage 8, resolutions regarding future behavior are largely determined by the affective tone of postoffense appraisals. This suggested model is preliminary and needs further validation. It also needs extension with respect to transitional mediating processes. 18 references