Despite recent advances in sex offender assessment, particularly in risk estimation and assessment of psychopathy, there are still gaps in what is known about sex offenders; for example, the actual prevalence rate for sex offenses is unknown, and research on juvenile and female sex offenders is sparse compared with that on adult male sex offenders. For a variety of reasons, problems in securing samples of sex offenders for research are virtually insurmountable. A different research methodology is now emerging that bases offender studies on the use of the polygraph. This research has found a significant amount of diversity in the targeting of victim types. Currently, there are no comprehensive theories on sex offenses or sex offenders, which may explain the broad-based approach often used in assessing sex offenders. This article discusses integrative approaches to understanding sex offending, testing procedures, the role of psychopathy, the role of actuarial risk assessment, and the theory behind sex offender assessment. In describing current sex offender assessment, the author notes a reliance on a multimethod assessment battery that involves test scores combined with history, information and data from other records, interviews with the offenders, the results of physiological measures, and the estimate of risk from one or more actuarial instruments. The process is strongly influenced by the theories underlying the tests used, along with the theoretical model of sexual violence held by the examiner. Lacking a comprehensive theory of sexual violence, clinicians rely on idiosyncratic combinations of data to develop a profile of the subject, including proclivities, weaknesses, strengths, and potential for recidivism. The author suggests that if future research incorporates nonlinear models with linear models, along with the statistical models now used, then a more accurate and useful understanding of sexual violence could emerge. 110 references