- Long-Term Follow-Up Studies Are Difficult:Comment on Langevin et al. (2004)
Understanding the long-term recidivism rates of sexual offenders is an important research topic with considerable implications for criminal justice policies. Estimating long-term recidivism rates is difficult, however, because the behaviour of interest is hidden by the offenders and infrequently reported by the victims. Researchers can only count offending that appears in the available records. As Dr Langevin is acutely aware, the correspondence between existing records and actual offending is far from ideal. Many cases do not result in convictions, and many convictions do not appear on the databases available to researchers. How researchers address these lacunae will influence the results they obtain.
It appears that the way in which Langevin, Curnoe, Fedoroff, Bennett, Langevin, Peever, Pettica, and Sandhu (2004) dealt with missing data inflated the observed recidivism rates of their sample. In an earlier report of the same data set, Langevin and Fedoroff (2000) reported that they were able to obtain follow-up criminal history records for 378 (54%) of the first 700 cases assessed at Dr Langevin's clinic between 1969 and 1974. In the 2004 report, the offenders lacking criminal history records in 1994 and 1999 were eliminated from the sample. Such a decision would retain recidivists and eliminate non-recidivists. The RCMP policies for retaining criminal history records have changed through the years, but they have always been more inclined to keep the records of active offenders (recidivists) than to keep those of inactive offenders (non-recidivists). In 1993, for example, the purge policy was that all convictions would be removed from the RCMP database if the offender (a) died, (b) was pardoned, or (c) attained age 70 and had no criminal activity for 10 years, or if (d) there was no recorded criminal activity for 15 years, provided that the offender was not the subject of a current investigation (Purge Unit, RCMP 1993). Given that the follow-up period in Langevin's study was more than 15 years, all of the inactive (non-recidivist) offenders should have been deleted from the RCMP database. [End Page 103]
By selecting offenders from the 1960s with active criminal history records in the 1990s, researchers should expect to find recidivism rates close to 100%.
Langevin et al. (2004) used an unusual definition of recidivism that includes both prior offences and future offences. In the earlier report (Langevin and Fedoroff 2000), the results are presented in a way that allows the reader to separate future offences (the usual definition of recidivism) from past offences. In the 378 cases with follow-up information, 139 (36.8%) had subsequent convictions for sexual offences. This figure is very similar to the long-term sexual recidivism rate (35.1%) found for another group of sexual offenders (child molesters) from Ontario in the 1960s and 1970s followed using RCMP records (Hanson, Scott, and Steffy 1995).
In my own work, I have become increasingly aware of the difficulty of estimating long-term recidivism rates from retrospective studies. It is not uncommon for records available in one follow-up period to disappear in subsequent years. With advances in computerization, the records should be getting better. Nevertheless, it is unrealistic to expect that complete criminal history records of all offenders will be retained in perpetuity by anyone. Researchers wishing to estimate lifetime recidivism rates require prospective studies – and patience.
1. The views expressed are those of the author and are not necessarily those of Public Safety and Emergency Preparedness Canada.