Once a sex offender, always a sex offender?

Currently, there is little empirical evidence to suggest a length of time at which a sexual offender is at risk of recidivism. I investigate whether it is possible to identify when sexual offenders can be considered low risk and theoretically pose no more of a threat than the non-offending population. I will be examining data from the Offenders Index, which contains the criminal histories of convicted sexual offenders from England and Wales from 1963 up to 2008. Eight birth cohorts, 1953, 1958, 1963, 1968, 1973, 1978, 1983 and 1988, will be analysed and compared.

In addition, I wish to determine whether criminal history and demographic variables increase or reduce a sex offender’s risk of re-conviction. Offender’s criminal histories are obtainable from the Offenders Index, however demographic variables are not. As a result, I have obtained access to Norway’s rich data which will allow me to see potential demographic risk factors, some of which are time-constant (e.g. country of birth and parents educational level at 16) and others time-varying (e.g. resident in Norway, employment, income and number of children). Norwegian conviction data is held from 1984 to 2009.
Although there are difficulties in comparing data across countries, especially crime data, studying the results of demographic variables on the risk of sexual conviction in Norway will help to shed light on the risk of sexual offending.
Survival analysis will be used to analyse the data, three statistical methods will be implemented; life-table analysis, Cox analysis and discrete time survival analysis with a complementary log-log link.

These methods will be used to estimate hazard curves for various types of offender and non-offender, and to see if and when they converge. Thus, empirical evidence will be obtained on the length of time at which a convicted sexual offender is no more of a risk of committing a further sexual offence than the non-offending population.