The current study has matched 936 of those released from Peterborough with 9,360 released from
other prisons on 36 out of 38 variables. While it was not possible to precisely replicate the approach
agreed in Cave et al. (2012) because of data quality issues, the Independent Assessors have
concluded that the model is sufficiently accurate and robust to support the reconviction analysis.
Furthermore, this analysis has shown that despite a 8.39% reduction in the number of reconvictions
in Cohort 1 when compared to the Comparison Group, the reduction was insufficient with regard to
the terms set out in the contract between the Ministry of Justice and the Social Impact Bond
partnership to be considered an Outcome, thus did not trigger payment.