04/07/2026
By Karen Mullins

The School of Criminology and Justice Studies is proud to announce a Dissertation Proposal Defense by Luke DeZago entitled, "Classifying Public Support for Police Use of Force: A Latent Class Analysis and Survey Experimental Analysis of US Attitudes, 1972-2026."


Date: Thursday, April 16, 2026
Time: 10:30 a.m. – noon
Location: Health and Social Sciences Building room 431, South Campus and via Zoom

Committee:

  • Jason Rydberg, School of Criminology and Justice Studies, UML (Chair)
  • Christopher Harris, School of Criminology and Justice Studies, UML
  • Kelly Socia, School of Criminology and Justice Studies, UML
  • Yan Wang, Department of Psychology, UML

Abstract:

Public support for police use of force (PUoF) is a consequential yet insufficiently theorized dimension of American public opinion. Although closely tied to police legitimacy, procedural justice, and broader punitive sentiment, prior research has largely relied on additive scales, individual sociodemographic correlates, and legacy survey items whose contemporary validity remains uncertain. This dissertation advances criminological public opinion research by reconceptualizing support for PUoF as a heterogeneous and potentially typological phenomenon rather than a single continuous disposition. Drawing on General Social Survey data from 1972 to 2024, linked with Uniform Crime Report and World Bank indicators, the study uses latent class analysis to identify distinct configurations of public support for PUoF and examines how these typologies relate to broader punitive attitudes, political orientations, psychological dispositions, and structural conditions. These analyses are extended through Bayesian hierarchical age-period-cohort characteristic models to assess historical and generational change, and the stability of these relationships across eras of policing reform. To address a longstanding measurement gap in criminology, the dissertation also incorporates an original randomized survey experiment comparing legacy GSS PUoF items with updated wording to evaluate whether widely used indicators continue to validly capture contemporary attitudes. By integrating person-centered classification, Bayesian multilevel modeling, and survey-based measurement validation, this project makes both substantive and methodological contributions to quantitative criminology. In doing so, it improves how scholars conceptualize, measure, and model punitiveness, while offering insight into the moral and political foundation of public support for coercive state power.