Overall, the evidence for program effectiveness, particularly in relation to cognitive-behavioural approaches (which typically incorporate a gendered analysis), appears to be more robust for sex offender intervention programs than domestic violence perpetrator intervention programs. At the same time, it is important to note that larger number of studies have evaluated the effectiveness of intervention programs based on CBT compared to programs based on other approaches. Thus, it is possible that, rather than being less effective or ineffective, non-cognitive behavioural approaches, have not yet received the empirical investment required to demonstrate their effectiveness (for example, owing to the fact that they less readily lend themselves to experimental or quasi-experimental designs).
The question of ‘what works for whom’ remains largely unclear, and research is still at an early stage in terms of identifying what program components are effective (see also Blatch & Delaney, 2009).
Recent evidence points to the need to match programs to perpetrator characteristics (e.g., risk, motivation, need), an important part of which is the initial screening and assessment process and the need for reliable and valid screening and assessment tools. On the whole, there appears to be a stronger correlation between socio-demographic variables and recidivism for domestic violence than sexual offender intervention programs. On the other hand, there appears to be a stronger correlation between psychopathology (e.g., personality disorders) and offender outcomes (i.e., recidivism) for sexual than domestic violence perpetrator intervention programs.
While there may be a need for intervention programs to pay greater attention to individual differences in risk, motivation and need, current programs do not typically address socio-demographic differences among perpetrators such as socio-economic status, ethnicity and geographic location.
It is worth noting that the emphasis on an “individualised” approach to intervention, as discussed in the literature, relates to matching the type of intervention to individual factors such as risk level (e.g., previous criminal history), cognitive capacity, and comorbid conditions (e.g., mental health and drug use issues). While an “individualised” approach is historically linked to a clinical or psychopathological perspective on domestic and sexual violence, the conceptualisation of an individualised or matched intervention approach does not preclude the adoption of a feminist perspective or the use of group-based interventions.
Contemporary psychological perspectives on human behaviour recognise that a multitude of factors influence individual attitudes and behaviour, and that these factors occur at different levels and have both independent as well as interactive effects on behaviour. People define themselves at multiple (e.g., individual, interpersonal and group levels), and conceivably, socio-structural factors such as gender power relations can be targeted at the individual as well as at the group and societal levels.
It is noteworthy that research into domestic violence perpetrator intervention programs is increasingly emphasising the risk, need and responsivity principles that have been more widely adopted in sex offender intervention programs. However, it is unclear the extent to which findings from sex offender programs are applicable to the domestic violence field. Given the importance of contextual factors on program outcomes, it is also unclear the extent to which overseas research findings are applicable to the Australian context.
Efforts that occur at the intervention program level should be accompanied by intervention at the broader societal level to address socio-structural factors that reinforce or perpetuate domestic and sexual violence. A multi-level approach is most likely to result in sustained outcomes.
Although current empirical evidence for program effectiveness is not considered to be robust for either domestic or sexual violence perpetrator intervention programs, it does not mean that current programs are ineffective in reducing recidivism or the economic and social costs to the broader society. As highlighted in this literature review, a number of program and non-program (e.g., research methodological, contextual) factors influence the capacity of empirical studies to demonstrate program effectiveness. A greater investment in research is therefore needed to disentangle program effects from other factors that can impact on program effectiveness and to determine the combination of strategies that maximise program effectiveness in a given context.
Research on intervention programs for domestic and sexual violence perpetrators remains at an early stage of development. Despite the progress that has been made, there is still a large gap in knowledge about how theoretical frameworks such as the feminist perspective on gender inequality should be addressed in intervention programs, and how intervention programs should be designed, delivered and examined empirically (e.g., Eckhardt et al., 2006). There is, however, a strong consensus among researchers that more and high quality evaluation studies of offender intervention programs are needed.
Researchers (e.g., Gondolf, 2004; 2009) are increasingly highlighting the importance of contextual factors in the effectiveness of intervention programs. The existing body of research, which is largely based in North America (USA and Canada), provides useful information about the design of intervention programs and their impact on psychological and behavioural outcomes. It is important, however, to assess the effectiveness of intervention programs in the Australian context. It would be important to conduct thorough evaluations of perpetrator programs in Australian (across geographic locations) to determine the extent to which findings from international research are applicable to the Australian context.
Central to any future research agenda is the need for adequate funding and resources, and collaboration between funders, researchers and service providers (Eckhardt et al., 2006). The Campbell Collaboration review also raised the importance of using larger samples, longer follow-up period, and both official and victim reports in the measurement of recidivism (Feder et al., 2008). Quality evaluation will also be supported by having stronger program logic (i.e., clearly articulated links among program objectives, activities, outputs and outcomes) underpinning intervention programs.
Humphreys, Ross and Diemer (2010) made a number of recommendations for future evaluations of men’s behaviour change programs. These recommendations included:
Information on the perpetrator’s history of offending, conditions of the intervention order and police risk assessment be available for program facilitators and evaluators
Systematic recording of non-attendance and sanctions and the provision of these data to the evaluators
Compulsory participation in evaluation is needed to ensure valid and reliable evaluation findings
Ability to observe program delivery practices
Inclusion of partner follow-up as part of the evaluation (by systematically collecting data on all partner contacts made or attempted and ensure that the data is available to evaluators)