The author first reviews relevant research that is hypothesis generating, in that it was not designed to test a model of overall sexual offender recidivism risk; however, the conclusions of such research are consistent. As a basis for generating hypotheses about the appropriate model for assessing recidivism risk, the research suggests there is evidence for more than one avenue by which recidivism risk can be driven, and two of the potential pathways are related to sexual deviance and general criminality/violence. This multidimensional model for sexual offender risk assessment already apparently has substantial empirical support. The most direct attempt to assess whether there are multiple underlying dimensions for recidivism risk is through factor analysis. There have been two factor analytical studies of risk considerations to date. The research by Roberts, Doren, and Thornton (2002) included two factor analytic studies. The first, involving 103 participants with 9 risk measures resulted in a 2-factor outcome: antisocial violence and sexual deviance/repetitiveness. The second factor analysis used a different set of 10 variables commonly shown to correlate with sexual offender recidivism. Three factors were identified: general criminality, sexual deviance, and detachment. In another study, Butz-Whittaker, Straussburg, and the Center for Family Development (2001) investigated the relative utility of static and dynamic sexual risk variables. The dynamic variables were classified as psychopathic (lacks remorse and empathy and is grandiose and superficial); resistant to treatment (lack of knowledge, ability, and motivation; lack of commitment to change; lack of goals; and poor judgment); and impulsive (poor behavioral control). Implications are drawn for research and application. 39 references