Association for Public Policy Analysis & Management

School safety has garnered much attention from policymakers and the public in recent years.  Despite this increased attention, the conversation has focused almost exclusively on the safety of students, leaving the safety of other key school personnel, such as teachers, unexamined (Espelage et al., 2013).  The available evidence suggests, however, that teacher safety is a salient component of the teacher experience and may be an important consideration for educational policymakers.  Recent evidence from a nationally representative sample of teachers revealed that over 80% of teachers have experienced some form of victimization (McMahon et al., 2014).  Teacher victimization may contribute to increased teacher turnover, with teachers either leaving the school or the profession.  Teacher turnover, in turn, is linked to lower student achievement (Ronfeldt, Loeb, & Wykoff, 2013).  The purpose of this study is to explore the relationship between teacher victimization and teachers’ decisions to remain teaching in their school or to remain in the teaching profession.  The one study that has addressed this issue finds potential relationships; however, the study does not distinguish between different types of victimization and fails to control for potentially confounding school level variables (Zurawiecki, 2013).  Utilizing data from multiple iterations of the nationally representative Schools and Staffing Survey (SASS), we address these gaps in the literature by examining the extent to which being threatened or attacked at school predicts higher rates of teacher turnover and whether this relation differs due to indicators of teacher resiliency.  We utilize ordinary least squares regression models with a robust set of controls, school fixed effect models to implicitly control for unobservable school factors, and propensity score matching.  We also explore the moderating effect of teacher resiliency, a combination of person and contextual protective factors, is explored as well as differential effects for teachers of different grade levels and by teacher subject field.  Findings from the OLS models suggest that experiencing an assault or threat at school predicts turnover from the school or profession with differential impacts of threats/assaults on turnover from the school and turnover from the profession.  This study expands upon prior research in the field by disaggregating effects by type of victimization (threat and assault), utilizing a significantly larger dataset with multiple time points, and including school fixed effects.  In doing so, this work contributes to a nascent body of literature on teacher victimization and informs an important policy lever by which teacher retention and student achievement may be improved.