Foundational Statistics for Risk Assessment
New to GIFR? Welcome! Create your account by clicking here. If you already have an account, please log in to proceed.
Training Description
In recent years, the use of structured risk assessment tools to assess recidivism risk has been increasing in criminal justice systems, mental health services, and the interface between them. Supporting the use of such instruments is a large evidence base that has grown exponentially in recent years. The aim of this comprehensive training is to provide attendees with the statistical knowledge necessary to be thoughtful consumers of peer-reviewed research studies on risk assessment tools so as to be able to apply study findings into practice. First, descriptive and analytic epidemiological principles will be described and their application to the field of forensic risk assessment explored. Second, null hypothesis statistical testing will be explained. Third, parametric and non-parametric statistical tests of differences will be reviewed, as will different approaches to correlation and regression. Fourth, the differences between discrimination and calibration effect sizes used to measure the accuracy will be discussed. Fifth, common statistical criticisms regarding the use of risk assessment tools will be presented. Sixth, statistics used to measure the inter-rater reliability and internal consistency of risk assessment tools will be identified. Finally, differences between the four main types of reviewing will be examined (narrative reviews, systematic reviews, meta-analyses, metareviews). Practical examples and illustrative videos will be used to illustrate key points.
Trainer Biography
Jay P. Singh, PhD, PhD is a Fulbright Scholar, Clinical Associate at the University of Pennsylvania, and Visiting Scholar at the University of Cambridge. Former Senior Clinical Researcher in Forensic Psychiatry and Psychology for the Department of Justice of Switzerland and author of over 75 peer-reviewed publications and books, he completed his graduate doctoral studies in psychiatry at the University of Oxford and clinical psychology at Universitat Konstanz. Since this time, he has lectured for Harvard, Yale, Oxford, Cambridge, Columbia, Cornell, Brown, Dartmouth, and UPenn. He was named the youngest tenured Full Professor in Norway in 2014 and is currently affiliated with the University of Pennsylvania and the University of Cambridge. A popular guest on both national and local television and radio, Dr. Singh's work has been featured in leading newspapers such as The Washington Post and magazines such as People. He has been the recipient of awards from organizations including APA Divisions 41 and 52, the Royal College of Psychiatrists, the European Congress on Violence in Clinical Psychiatry, the Society for Research in Child Development, and the Society for Research in Adolescence.
Learning Objectives
This training is designed to help you:
- Examine the difference between descriptive epidemiology and analytic epidemiology
- Describe strengths and weaknesses of effect sizes most commonly reported in risk assessment tool studies
- Discuss how null hypothesis statistical testing is used in research studies
- Differentiate between parametric and non-parametric statistics
- Discuss how risk assessment tools are statistically developed
- Differentiate between discrimination and calibration effect sizes
- Examine common statistical criticisms regarding the use of risk assessment tools
- Identify the four major types of reviews used to synthesize research
