Cognitive-behavioral therapy (CBT) is among the more promising rehabilitative treatments for
criminal offenders. Reviews of the comparative effectiveness of different treatment approaches
have generally ranked it in the top tier with regard to effects on recidivism (e.g., Andrews et al.,
1990; Lipsey & Wilson, 1998). It has a well-developed theoretical basis that explicitly targets
“criminal thinking” as a contributing factor to deviant behavior (Beck, 1999; Walters, 1990;
Yochelson & Samenow, 1976). And, it can be adapted to a range of juvenile and adult
offenders, delivered in institutional or community settings by mental health specialists or
paraprofessionals, and administered as part of a multifaceted program or as a stand-alone
intervention. Meta-analysis has consistently indicated that CBT, on average, has significant
positive effects on recidivism. However, there is also significant variation across studies in the
size of those treatment effects. Identification of the moderator variables that describe the study
characteristics associated with larger and smaller effects can further develop our understanding
of the effectiveness of CBT with offenders. Of particular importance is the role such moderator
analysis can play in ascertaining which variants of CBT are most effective. The objective of this
systematic review is to examine the relationships of selected moderator variables to the effects
of CBT on the recidivism of general offender populations.