Using Quasi-experimental Methods to Understand the Effects of School Security Investments on Schools and Students

• In Event: Methodological Pluralism for Comprehensive Evaluation of the Implementation and Impacts of School Security Equipment

Abstract

Objectives
Rising concerns over school shootings and violence have fueled a multi-billion-dollar school security industry, with over $3 billion spent annually in the U.S. on security infrastructure. This spending rests on the theory that increasing physical security will reduce violence and improve safety. Yet existing research provides limited support for this assumption and is primarily correlational.

This study is the first, to our knowledge, to use quasi-experimental methods to estimate the causal effects of school security investments on outcomes related to safety, school climate, discipline, academics, and juvenile justice involvement. Using data from Virginia and a dynamic regression discontinuity design, we examine the impact of receiving a school security equipment grant on both school-level and student-level outcomes, with particular attention to differential effects for marginalized student populations.

Conceptual framework
We draw on theories of securitization and racialized social control, particularly those related to the school-to-prison pipeline. These frameworks challenge the assumption that increased surveillance and policing in schools lead to greater safety. Instead, they highlight how such measures may disproportionately criminalize students of color and those from other historically marginalized groups, reshaping educational spaces into sites of control rather than protection.

Methods
We implement a fuzzy dynamic regression discontinuity design, adapted from Cellini, Ferreira, and Rothstein (2010). In the first stage, we model cumulative grant receipts per pupil using instrumental variables based on award cutoffs and application history, as well as polynomial functions of prior application scores and fixed effects for school and year. In the second stage, predicted grant amounts from the first stage are used to estimate effects on a range of school- and student-level outcomes.

Data
The study draws on ten years (2014–2023) of school security grant data from the Virginia Department of Education. These records are linked to statewide student-level administrative data and juvenile justice records. We supplement these with publicly available school-level data on discipline, climate, and other safety-related indicators.

Results
We have validated the discontinuity in the grant scoring process and confirmed the feasibility of our identification strategy. Simulations and preliminary analyses suggest the design is well-powered to detect unbiased impacts on school-level safety outcomes. Full analyses are underway and will include both aggregate and subgroup-specific estimates.

Significance
As school districts continue to invest heavily in security infrastructure, rigorous evidence on the actual impacts of such investments is urgently needed. This study advances knowledge in three important ways. First, it moves beyond correlational evidence to provide causal estimates of school security's effects. Second, it evaluates a wide range of outcomes, recognizing that the consequences of security equipment may extend beyond violence prevention to affect school climate, discipline, academic achievement, and youth justice involvement. Finally, it attends to equity by examining whether these impacts differ for marginalized students. Findings will inform policymakers weighing the trade-offs of security-focused school safety strategies.

Authors

• Lucy Sorensen, University at Albany - SUNY

• Samantha L. Viano, George Mason University

• Antionette Stroter