AbstractThe goal of this retrospective, correlational study was to determine if the MTC:CM3 classification system yields any interrater reliability and validity for sample of pedo-sexual offenders from an outpatient population. Three independent raters classified each 124 subjects, by content analysing archival data such as clinical and judicial case files. The data was also coded on the basis of psychometric and demographic data, such as self report measures on cognitive distortions, personality traits, etc. and more general characteristics like age, number of victims, marital status, etc. We used our Dutch translation of the Massachusetts Treatment Center: Child Molester Typology (3rd revision), operationalising the original criteria by Knight & Prentky (1989) in a checklist format.
Two kinds of multiple interrater reliability (IRR) analysis were conducted: we differentiated between the IRRs before and after a joint deliberation of the ratings. For the validation analysis several parametric and non parametric techniques were used such as Pearsons Chi-Square, Fisher Exact test, one-way ANOVAs, independent-samples t-tests, Mann-Whitney or Kruskall-Wallis, depending on the respective hypothesis and variable.
The results of IRR analysis before deliberation were acceptable to satisfactory with Cohens Kappa values varying between .49 and .74., whereas the IRRs after deliberation clearly improved to Kappa scores ranging between .89 and 93. which seems a good indication of the replicability of the system in a different population.
As for the validation analysis we found strong support for only several categories of the MTC:CM3, but none for others. The most two common types of pedosexual offenders are not sufficiently differentiated by the classification system. Consequently we propose several adjustments and recommendations for further research.
|Date of Award||21 Sep 2004|
|Supervisor||Hedwig Sloore (Promotor), Christiaan Schotte (Jury) & Anne Brewaeys (Jury)|
- sex offending
- MTC:CM3 typology