Careers of institutionalised serious offenders

Criminology Research Council grant ; (3/88)

The first report, 'Careers' of Chronic Institutionalised and Serious Offenders, takes two approaches to the question' do offenders tend to commit more serious crimes with continuing returns to prison?' An answer to this question was sought by analysis of a large data file which is described in the report. The first approach, using a 'one-step' method of estimation of probabilities of recidivism on successive returns to prison, showed overall probabilities of recidivism rise rapidly on successive returns to prison, plateauing at about 0.8 (for male non-Aboriginals) and at about 0.9 (for male and female Aboriginals) by about the fourth or fifth imprisonment. The numbers for female non-Aboriginals were too small for estimation. At each step, roughly equal proportions were estimated to be imprisoned for offences which were less, equal, or more serious than the previous, according to an offence rating scale devised in the report. There were some small but significant departures from equality of proportions. The second approach looked along an offender's record and asked whether the offender ever committed a crime more serious than the first and, if so, whether the offender subsequently ever committed a more serious crime.

The second report restricts itself to offenders convicted at, some time in their 'careers' of the following types of offences:

  • carnal knowledge and other sex offences excluding rape;
  • rape; and
  • homicide, wilful murder and attempted murder.

The subjects in the study were derived from a large computerised prisoner record file of offenders released for the first time from Western Australian gaols between 1975-1987.

The study addressed the question of the probabilities of return to prison for any offence, subsequent to an offence of the above type, and how it varied with race, type of offence and whether prior imprisonments had been recorded. In addition, it addressed the probabilities of imprisonment for another violent offence at any time subsequent to imprisonment for an offence of the above type, and how it varied with the above factors.

In summary, the research found:

  • A relatively high proportion of offenders eventually repeat the same or similar offences, especially those involving sexual assault.
  • Serious sex offenders and non-Aboriginals had approximately a 38 per cent chance of returning to prison for any offence, while Aboriginals had a 77 per cent chance. Having a prior record significantly increased the probability of failing for either race.
  • High probabilities were also estimated for the sex offenders (homicide cases excluded) to return for another violent of- fence. In excess of 20 per cent of non-Aboriginal cases were estimated to do so, and for Aboriginals the estimate exceeded 60 per cent. Prior record influenced rapidity of failure (those with prior records failed sooner) but not the probability of ultimate repetition of a violent offence. Small numbers of cases presented some problems in these analyses.
  • For homicide offenders probabilities of failing for any offence did not differ from those of sexual offenders. There were too few cases to allow estimation of failure to another violent offence.
  • Detailed consideration of the records of these offenders suggested that recidivists tend to be 'generalists' rather than 'specialists' in offence preferences. But prisoners incarcerated for incest provide little evidence of prior violent or sexual offences.
  • Overwhelmingly, violent offences seem associated with offenders who show a pattern of aggressive behaviour, suggesting that aggression rather than perversion may be the more salient characteristic of sexual assault offenders. Nevertheless large numbers of sex offenders never return to prison.