Chapter 2 The development of OASys Social/dynamic factors The reason prediction scales have tended to be based on criminal history variables is that the data is readily available and considered to be accurate and objective. There are many other social or dynamic factors that could be used relating to the offender’s personal situation. They are ‘dynamic’ because they can change over time, unlike criminal history. For example, an offender can become more or less drug dependent, or they can obtain or lose employment. Data relating to social dynamic factors is far more difficult to obtain because this information has never been collected consistently in the UK.
Dynamic factors are considered to be more difficult to measure reliably than criminal history variables because information has to be compiled from many different sources, and often requires interpretation. Two definitive research reviews, Hollin and Palmer (1994), and Gendreau and Coggin (1996), have summarised the research using social dynamic factors to predict re-offending. Both reviews conclude that social dynamic factors are predictive of re-offending. Indeed, Gendreau and Coggin concluded that social variables can be as predictive as the criminal history data.
A meta analytical study (Gendreau, Little & Coggin, 1996) compared the results of 131 published studies on the prediction of recidivism in adult male offenders and found that the best predictors of recidivism were first, ‘attitudes and behaviours supportive of a criminal lifestyle’, second, criminal history, and third, social achievement (employment or education). Most of the studies cited in these reviews are North American but there is some recent British work. May (1999) conducted a large-scale British study to examine the link between social factors and reconviction for offenders given a community penalty.
The study aimed to establish whether including social factors along with criminal history variables could improve the prediction of re-offending rates. May concluded that the most powerful predictor of reconviction was criminal history, but that social variables can play a part, particularly for those with multiple problems and for less experienced offenders. Of the social variables considered in the study, drug problems were most strongly linked with reconviction, and problems with employment, accommodation and finances were also significant. This major study, which involved a sample of 7000 offenders, was hampered by the lack of information available on offenders’ social circumstances.
The author recommended that social factors would be most useful for prediction if collected fully and consistently with a single national approach to recording. A study by Clark (1998) showed that social factors were important in predicting reconviction for a group of about 400 prisoners following discharge. In this study, education and employment history improved the level of accuracy of prediction achieved by criminal history variables alone. Measures of underlying personality traits, patterns of behaviour and OASys Manual V2 © Crown Copyright 2002 Page 7