Date of Award
Spring 2026
Language
English
Embargo Period
4-30-2028
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
College/School/Department
Department of Public Administration and Policy
Program
Public Administration and Policy
First Advisor
Lucy C. Sorensen
Second Advisor
Shawn D. Bushway
Committee Members
Matt Vogel
Keywords
school safety, school resource officers, school shootings
Subject Categories
Education Policy
Abstract
Positive school climates promote better educational and health outcomes for students and are thus critical for their success and long-term wellbeing. School safety, which encompasses the physical and emotional security of a school community, is a key dimension of school climate and has emerged as a recent policy priority in the United States. Creating safe schools involves not only identifying the risk factors that contribute to violence but also ensuring that interventions aimed at reducing such violence do not harm student wellbeing. By addressing these two elements of school safety, this dissertation enhances our understanding of effective school safety practices that foster a positive school climate for all students.
Informed by an ecological framework of school violence, the first paper uses school and community level data to investigate predictors of school shootings and explore how such predictors can be used to better allocate prevention resources across schools. To capture relevant school and community characteristics, I merge data from multiple sources including the K-12 School Shooting Database, the Civil Rights Data Collection, the US Department of Education’s National Center for Education Statistics Common Core of Data, the American Community Survey Education Tabulation Survey Data, the FBI’s Uniform Crime Reporting Data, and the US Census Bureau’s Annual Estimates of Resident Population for Counties. Using three predictive machine learning algorithms, including logistic regression, random forest, and gradient boosting, I find that six school characteristics are consistently associated with a higher risk of a school shooting occurrence. These characteristics include student enrollment, percentage of White students, percentage of students with disabilities, school level, out-of-school suspensions, and the number of counselors per 100 students. These findings suggest how the complexity of socioeconomic contexts is associated with the risk of school shootings. To illustrate how these factors can be used to target schools when allocating prevention resources, I create a risk-informed targeting strategy and compare it to other hypothetical strategies. The exercise shows that the risk-informed strategy better concentrates prevention resources among schools with a high risk of shooting. This finding suggests that using a composite risk score based on observable school characteristics can complement existing funding mechanisms and enhance the efficiency of resource allocation.
The second paper consists of two parts. The first part reviews and summarizes existing legislation across 29 states regarding requirements for specialized training that prepares school resource officers (SROs) to work in a school setting. Findings show that training mandates vary greatly across states, with some states providing more detailed training items than others. However, decreasing the involvement of children and adolescents with the criminal justice system remains a common motivation underlying training requirements across most states. Furthermore, the review identifies the cases of two states, Colorado and Texas, which are then analyzed in greater detail in the second part of the paper. The second part thus focuses on exploring the effects of training mandates on the use of disciplinary practices, school-based arrests, and referrals to law enforcement. To answer this question, I use two waves of data from the Civil Rights Data Collection and a triple differences methodology to separately analyze the cases of Colorado and Texas. I find different impacts across states. In Colorado, training mandates are associated with an increase in the use of out-of-school suspensions and expulsions and a decrease in the use of in-school suspensions. Findings suggest that there is an escalation in the severity of disciplinary responses to student misbehavior but without the need to further involve law enforcement. In Texas, training mandates are associated with an increase in student arrests and referrals to law enforcement, as well as a decrease in harassment and bullying. Findings suggest that law enforcement actions did not replace school discipline practices but instead layered on top of them, potentially amplifying punitive practices. Additional analyses by student subgroups for both states suggest that mandates affected students differently. Potential explanations behind the different impacts of training observed in Colorado and Texas include differences in the training content required by each state. Overall, findings highlight the importance of clear goals and measurable outcomes of training mandates moving forward.
By investigating characteristics of the school climate that may be more conducive to school shootings, one of the most extreme forms of school violence, and exploring the consequences of training mandates for a school safety intervention, this dissertation advances our understanding of school safety and complements policy efforts to promote safe and supportive school climates.
License
This work is licensed under the University at Albany Standard Author Agreement.
Recommended Citation
Avila Acosta, Montserrat, "Toward Safer Schools: Empirical Essays on School Shooting Risk and School Resource Officer Training" (2026). Electronic Theses & Dissertations (2024 - present). 431.
https://scholarsarchive.library.albany.edu/etd/431