Recidivism Among Child Sexual Abusers: Initial Results of a 13-Year Longitudinal Random Sample

Abstract

In the initial analysis of data from a random sample of all those charged with child sexual abuse in Idaho over a 13-year period, only one predictive variable was found that related to recidivism of those convicted. Variables such as ethnicity, relationship, gender, and age differences did not show a significant or even large association with recidivism. The only variable that seemed to show both a significant and almost moderate association to recidivism was the Risk Assessment in the Sex Offender Evaluation reoffense. Comparisons were made to prior research as well as a discussion of implications of the sex offender evaluation for the legal process. Finally, a call for the continued need for further research is discussed.

KEYWORDS:

• child sexual abuse
• recidivism
• longitudinal analysis

Notes

For the years 1991 through 2003, two years (1993 and 2000) are not included. For 1993 data, the authors were making the transition from mainframe analysis of the data to PC–based analysis; while the hard copy of the data exists, the computer-based data were lost in the transition. It was decided for this initial analysis that 1993 data would be excluded. There appear to be no significant differences between 1993 and the surrounding years. In 2000, the Idaho Supreme Court decided that the authors were no longer allowed access to presentence investigation reports for confidentiality reasons. For year 2000, social security numbers were not recorded, thus making it impossible to select a random sample for this year. Again, the year 2000 does not seem significantly different from the surrounding years. In conclusion, the exclusion of these two years, while regrettable, does not seem to affect the results of this analysis.

There are a large number of potentially complicated reasons data were not available for over 150 of the cases requested from the Department of Corrections. A significant reason is likely the random nature of the analysis. In each year's report, 10% to 15% of the charges filed are acquitted or dismissed. If these cases have no other convictions, there is no reason they would be in the records of IDOC. This means that 60–90 of the 600+ cases should not be in the records. Additional reasons may also be coding errors on the part of the authors, incorrect social security numbers in the records, and/or missing information in the electronic data recorded by the Department of Corrections. An analysis comparing the cases for which data is available and the missing data showed no significant differences.