School Violence Exposure as an Adverse Childhood Experience

Overview

This study hypothesize that school safety strategies may contribute to poor mental health and distress among various school populations and proposes the following research questions to be answered through this study: 1. Is there a significant difference in the prevalence of mental health outcomes, perceived school safety, and academic engagement between early adolescent and adolescent students (grades 6 - 12) at schools that have experienced a school shooting and those that have not? 2. Are existing interventions to promote school safety and security associated with poor mental health outcomes among students and school staff? 3. Does the strength of this association between school safety interventions and mental health outcomes differ among students and teachers in those schools who have experienced a school shooting versus among students and teachers in those schools who never experienced a school shooting? And, among our sample of students, are these associations moderated by their ACE history? Participants will be asked to self-report data on their mental health and wellness, and perceived school safety through a survey. Researchers will compare the surveys from the participants from exposure schools (those who experienced a gun shooting incident) with surveys from the participants from non-exposure schools (those who have not experienced a gun shooting incident)

The recent increase in school gun violence is an important cause of poor mental health and adverse childhood experiences (ACEs) among the school going populations. Schools have implemented a number of school safety and security measures, both in response and anticipation of gun violence. Some of these interventions increase anxiety, depression, and other indicators of poor mental health among students and staff alike. Despite this, the association between exposure to existing school safety interventions and early adolescent student mental health outcomes has yet to be investigated. This observational case-control study aims to understand this association by collecting cross-sectional surveys from students, teachers, staff, and principals across a sample of 12 nationally representative K-12 public schools in the United States. Data Collection Primary data collection- Among the participating schools, data will be collected through surveys prepared for students, teachers, and principals separately. Completing each survey will take about 25 minutes for the students, and 15-20 minutes for the teachers and principals. The student surveys will be administered by the research team in the classroom on a school day. Similarly, the teacher surveys will be administered on the school day. The principals' survey will be administered online via Qualtrics. These surveys will collect the following information from each participant: Demographics: Data on sex, race, age, grade-level, and presence at the time of the school shooting incident for students in the exposure school. Mental health and well-being: Among the participating students in each school, self-report data on mental health and well-being will be assessed via a survey that comprises 1) eight items from the Patient Health Questionnaire (PHQ) - A, a validated and widely used survey assessing mental health and symptoms of depression among youth; 2) 25 items from the Child and Adolescent Trauma Screen (CATS), which has been validated for use by youth ages 7-17 years; and 3) five items from the World Health Organization (WHO-5) Well-Being Index, which has been validated for use by individuals ages 9 years and older. The PHQ-A and CATS items are multiple choice with four response options, and the WHO-5 items are rated on a 6-point Likert scale. Among the participating teachers in each school, self-report survey data on mental health and well-being will be assessed via a) eight items from the PHQ - 9, a validated survey assessing mental well-being and symptoms of depression among adults, b) the Generalized Anxiety Disorder 7-item (GAD-7) scale, which has also been validated and assesses the most common anxiety disorders among adults, and c) via five items from the WHO-5 well-being index. The PHQ-9 and GAD-7 items are multiple choice with four response options and the WHO-5 items are rated on a 6-point Likert scale. ACE prevalence: Each student will provide self-report data on ACE prevalence. Ten of these items were adapted in previous research from a validated tool. Building upon additional research that has argued that children experience a greater range of adversity than most ACE assessments currently capture an additional five items will be included. The survey will therefore assess the following: bullying, residential instability, Child Protective Services (CPS) involvement, meeting of basic needs, parents divorced, death of a primary caregiver, family member with poor mental health, family member engaging in substance use, domestic violence, incarcerated parent, community violence, death of friend(s), experience with the foster care system, experience with the juvenile justice system, and exposure to violence in school. Each item will be multiple choice, with the following response options: "yes", "no", or "prefer not to answer" and these data will comprise an ACE prevalence score that denotes how many ACEs each student has experienced. Perceptions of school safety: The "Safe and Responsive Schools" Safe School Survey is a validated tool designed to be administered to students and school staff. This instrument assesses six critical constructs, including perceptions of school climate, school safety, and belongingness. In line with previous research on school violence and in an effort to comprehensively assess perceptions of school safety, this tool seeks to capture both minor conflicts (e.g., arguments among students) as well as more significant forms of disruption (for example, weapon possession on school grounds). Educational outcomes: Among the participating students in each school, self-report survey data on academic engagement will be assessed via the Student Engagement Instrument (SEI). The SEI is a 35-item survey that assesses multiple academic constructs, including extrinsic motivation to learn, future goals and aspirations, class participation, teacher-student relationships, and perceived peer support for learning. The survey also includes one multiple-choice item that has been adapted from the CDC's Youth Risk Behavioral Surveillance System (YRBSS) and that asks students to describe their academic grades in schools over the past 12 months. School safety and security strategies: Data on school safety and security strategies from the following seven categories will be collected- 1) external target hardening efforts (e.g. monitored school entry doors; signs indicating the school is a "Gun Free School Zone"), 2) internal target hardening efforts (e.g., metal detectors; security cameras); 3) student/staff monitoring (e.g. threat assessment team; zero-tolerance policies for weapons) 4) emergency procedures/drills (e.g., written active shooter plan); 5) emergency notification technologies (e.g., anonymous thread reporting systems; panic buttons); 6) medical support (e.g., full or part-time school nurse on campus); and 7) school security staff (e.g., law enforcement officers or police; teachers and/or other personnel armed with guns). This data will be collected in two ways- One, principals from all participating schools will be surveyed on their knowledge of current school safety tactics and policies. Second, the researchers will review each school's current school safety plan, which is a comprehensive document developed by school stakeholders that details preparation and response protocols to various school emergencies. Independent Covariates. Data on multiple independent covariates will also be collected via sources like the National Center for Education Statistics (NCES) and the U.S. Census and adjusted for in conditional logistic regression models. These covariates will include the length of time since the school experienced a shooting; demographics (e.g., urbanicity, school level, and type of school); and school district characteristics. In addition, secondary data on education access, equity, and school discipline will be directly downloaded from the Civil Rights Data Collection website. Using these recent data on each participating school, the study team will create a "school discipline" summative score that will comprise the following indicators (accounting for school population size and reflecting the number per academic year): attendance rate, number of in-school suspensions, out-of-school suspensions, students referred to law enforcement, students disciplined for harassment or bullying, and school-related arrests. Propensity score match and control statistical methods will be used to ensure rigor in association validity controlling for these characteristics comprehensively. Human Subjects Protection This study protocol has been approved by the Institutional Review Boards (IRB) at Columbia University Medical Center and Teachers College, Columbia University. The survey data collected from students, teachers, and school staff will be confidential, but not anonymous. All participants will be associated with a specific school. However, no identifying names or other identifiers will be used at any point in the study. Each participant will be identified solely via an alphanumeric code during the data collection and organization process. Data Organization and Treatment Survey data entry: All survey data collected in this study are being recorded initially in Microsoft Excel for organization and management and then subsequently read into SPSS (version 28.0) for analysis. The study team maintains a codebook; where needed, variables will be recoded and summative scores computed. Document review: Each school safety plan will be reviewed and coded independently by two members of our research team to determine if each of the school safety strategies and tactics of interest are being implemented. These data will also be recorded initially in Microsoft Excel for organization and management and then subsequently read into SPSS for analysis. Missing data: Consistent with previous research, we estimate the majority of survey items to have less than 10% missing data and no more than 19% missing survey data on any given item. However, should higher levels of missing data occur, multiple imputation methods will be used. Data Analysis Plan The data collected from the students, teachers, and principals across all 12 participating schools will be analyzed and will fully respond to each of the research questions. Descriptive statistics will be calculated to summarize student and teacher participant characteristics within each school and also describe each school's safety and security strategies. Analyses of Covariance (ANCOVAs) will subsequently be used to identify potential significant differences in the prevalence of mental health and well-being, perceptions of school safety, educational outcomes, and ACEs between students in schools who have more recently experienced an intentional school shooting (nAnticipatedRange(AR) = 375-600), schools who have experienced an intentional school shooting less recently (nAR = 375-600), and schools who have never experienced a shooting (nAR = 750-1200), while controlling for key covariates. Analyses of Variance (ANOVAs) will also be used to identify potential significant differences in the prevalence of mental health and perceptions of school safety between teachers in schools who have more recently experienced an intentional school shooting (nAR = 15-30), schools who have experienced an intentional school shooting less recently (nAR = 15-30), and schools who have never experienced an intentional school shooting (nAR = 30 -60). The research team utilized G\*Power (v.3.1) to run power analyses in anticipation of the proposed analyses, assuming an alpha-level of .05 and a power of 0.80. In the case of the proposed student analyses, a very small minimum effect size will need to be detected to meet the power requirements and given the anticipated minimum sample size. In the case of the teacher analyses, a medium effect size will need to be detected. Multiple regression analyses will then be used to determine if the strength of the association between school safety interventions and mental health outcomes differs among students and school staff in schools with varying levels of trauma exposure. Mediated moderation analyses will evaluate the role of ACEs on the relationship between exposure to an intentional school shooting, exposure to school safety strategies, and student outcomes (mental health and well-being, perceptions of school safety, and educational outcomes). Again, power analyses revealed that for the student-level analyses and given an alpha-level of .05 and a power of 0.80, a small effect size will need to be detected in order to meet the power and sample size requirements. In the case of the proposed teacher analyses, a medium effect size will need to be detected. Potential Study Challenges The researchers have identified some potential challenges this study may face and their possible solutions. Being a survey, the study requires voluntary participation from the students, teachers, and principals of the 12 selected schools. The participation will be encouraged by prioritizing confidentiality, reducing the time burden associated with the study, creating an accessible survey questionnaire, and creating an inclusive study environment by ensuring that all participants have regular opportunities to ask questions and express concerns throughout the study process. We will also be providing incentives to all participating schools to recognize their time spent on this study. Other potential risks include the possibility of emotional responses that may emerge during the data collection process. The research team will take great care to ensure that the proposed data collection effort takes place in a manner that is comfortable and safe for all participants. It should be noted that this study will not be collecting data on self-harm or suicidality. Finally, there is the possibility of missing survey data; however, the use of multiple data sources where possible and multiple imputation methods will address the problem of missing data during the data analysis process.