Accuracy of statistical models for crime linkage and the exploration of sexual offence classes

Kelkar, Gauri Milind ORCID: 0009-0002-4428-1601 (2024). Accuracy of statistical models for crime linkage and the exploration of sexual offence classes. University of Birmingham. Foren.Clin.Psy.D.

Kelkar2024ForenClinPsyD_Redacted.pdf
Text - Redacted Version
Available under License All rights reserved.
Download (1MB) | Preview

Abstract

Chapter one of this thesis is a meta-analysis of the accuracy of statistical models used to link crimes, as measured by the Area Under the Curve statistic. This chapter includes the process used to gather, and extract relevant papers and data, the model selection, and findings of a three factor random effects model. In the empirical research paper (chapter two), subgroups of offence behaviours, offender traits, and demographic characteristics are identified in sexual offences. This research is based on the Swiss ViCLAS database and uses latent class analyses to identify subgroups of different variable sets. Associations between the three sets of subgroups are also evaluated using log-linear analysis. Finally, chapter three and four present press release documents of chapter one and two respectively. These chapters are written for the purpose of public dissemination. They both include brief overviews of the aims, findings, and conclusions from the meta-analysis and empirical research chapters.

Type of Work:

Thesis (Doctorates > Foren.Clin.Psy.D.)

Award Type:

Doctorates > Foren.Clin.Psy.D.

Supervisor(s):