Date of Award
1-1-2019
Language
English
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
College/School/Department
Department of Sociology
Content Description
1 online resource (vi, 138 pages) : illustrations (some color)
Dissertation/Thesis Chair
Steven F Messner
Committee Members
Glenn Deane, Joanne M Kaufman, Alissa Worden
Keywords
Implicit Bias, Phenotype, Race/Ethnicity, Sentencing, Sex Offending, Discrimination in criminal justice administration, Sex offenders, Sentences (Criminal procedure)
Subject Categories
Criminology | Sociology
Abstract
The role of offender race/ethnicity and potential bias in criminal case outcomes is a popular topic both culturally and academically. Although a common research subject, the existing literature remains inconsistent and limited when focusing on sex offender sentencing outcomes. This dissertation uses data collected from the New York State public sex offender registry on white, black, and Hispanic males to examine the effect of offender racial/ethnic phenotype on two sentencing outcomes: sentence type and sentence length. Offender phenotype is measured through three facial features: nose width, lip fullness, and eye shape. These facial features were chosen from existing literature, however, I extend current methodological approaches by using a pre-trained machine learning algorithm to locate facial coordinates on images and then calculate the Euclidean distance between these coordinates to quantify facial features removed from human perception and bias.
Recommended Citation
Walsh, Christine M., "Stereotypes and phenotypes : using machine learning to examine racial implicit bias in sex offender criminal case processing" (2019). Legacy Theses & Dissertations (2009 - 2024). 2405.
https://scholarsarchive.library.albany.edu/legacy-etd/2405