Developing the actuarial prediction of violent and sexual reoffending

Howard, Philip (2014). Developing the actuarial prediction of violent and sexual reoffending. University of Birmingham. Ph.D.

PDF - Redacted Version

Download (1MB)


This thesis aims to develop and improve the actuarial prediction of violent and sexual offending. It demonstrates the importance of understanding offence classification and specialisation, and the value of dynamic risk factors in actuarial risk prediction. Its findings are especially relevant to prison and probation risk assessment and management practice in England and Wales, where the National Offender Management Service (NOMS) makes extensive use of the Offender Assessment System (OASys).
Chapter 1 takes a novel, empirical approach to determining which offences should be counted as “violent” by a new nonsexual violence risk scale. Chapter 2 then develops this new scale, the OASys Violence Predictor (OVP), which combines static and dynamic risk factors, and validates it through comparison with NOMS’s existing scales. Chapter 3 then shows that OVP is also an equally good or superior predictor of nonsexual violence among offenders with a history of sexual offending. Chapter 4 shows that OVP’s dynamic risk factors have causal properties and reassessment over time improves prediction. Chapter 5 demonstrates the significance of offence specialisation by sexual offenders to risk predictor development. Chapter 6 concludes the thesis with an overview and discussion of the findings, limitations, practical implications, and future research directions.

Type of Work: Thesis (Doctorates > Ph.D.)
Award Type: Doctorates > Ph.D.
College/Faculty: Colleges (2008 onwards) > College of Life & Environmental Sciences
School or Department: School of Psychology
Funders: None/not applicable
Subjects: B Philosophy. Psychology. Religion > BF Psychology
H Social Sciences > HN Social history and conditions. Social problems. Social reform


Request a Correction Request a Correction
View Item View Item


Downloads per month over past year