Sara Rahman, from the NSW Bureau of Crime Statistics and Research, delivered a talk on the “Evaluation of the Violent Offender Treatment Program” for the June SSA NSW Branch meeting. She is part of BOCSAR, a statistical and research agency within the Department of Justice, and conducts research focussed on crime and criminal justice.
Sara explained that the Violent Offender Treatment Program (VOTP) is an intervention offered to medium- to high-risk violent offenders and the main aim of her work is to evaluate the impact of VOTP on criminal justice outcomes. She carefully explained the various complexities in the process that make the evaluation challenging. This includes selection biases and confounding factors that may make the estimation of the effect of VOTP unreliable.
To properly estimate the effect of VOTP, first she recognised that the outcome variable was coded to consider that there may be censoring (the offender may be retained in custody and was not observed to offend by follow-up period) and re-offence may not necessary be violent. Then, she explained about using sentence length as a possible instrumental variable (IV). It was hypothesised that the sentence length may be correlated with time from referral to release thus the IV was residualised by sentence length. She showed further tests as evidence for the appropriateness of sentence length as IV. The effect of VOTP was measured by a two stage least squares approach. She concluded with the limitations and challenges in evaluating the effectiveness of VOTP.
Dr Emi Tanaka, NSW Branch Secretary