Background/Context: Staff turnover is a critical issue for practitioners and policymakers alike. Especially since the onset of the COVID-19 pandemic, significant attention has been dedicated to developing policy interventions to address high staff shortages. Intervention around staffing instability may be particularly pertinent for special education positions, as they are especially likely to be subject to turnover and vacancy (Nguyen et al., 2022). This includes both special education teachers and paraeducators (also known as paraprofessionals or instructional aides; Goldhaber et al., 2023; Theobald, Kaler, et al., 2023). Numerous policy interventions have been put in place to address teacher shortages and turnover (Theobald, Xu, et al., 2023; Zamarro et al., 2023), but we cannot say with certainty which types of staff turnover impact which students most significantly, and thus which policy interventions may be most pertinent. While extant research indicates that teacher turnover has a detrimental impact on student outcomes (Ronfeldt et al., 2013), we have no current estimates of the impact of paraeducator turnover, despite the fact that the size and growth of the paraeducator workforce outpaces that of teachers (Bisht et al., 2021; Theobald, Kaler et al., 2023). Moreover, there are currently no causal estimates -- to our knowledge -- of the impact of teacher or paraeducator turnover on students with disabilities. Purpose/Objective/Research Questions: This study seeks to address that gap in the literature by serving as the first, to our knowledge, causal estimate of special education staff turnover on outcomes for students with disabilities. Findings of this study will contribute to the literature by providing an estimate of how turnover across multiple staffing categories may have differential impacts across two key student groups. Additionally, our work is poised to complement recent studies on staffing instability, particularly within the special education workforce (e.g., Bettini et al., 2022; Theobald, Goldhaber, et al., 2021; Gilmour et al., 2022; Scott et al., 2023), with implications for educator workforce policy. This analysis will address the following questions: 1. What is the impact of paraeducator turnover on test scores for students with and without disabilities? 2. What is the impact of teacher turnover on test scores for students with and without disabilities? 3. How do these impacts differ by school characteristics? Setting: In this analysis, we leverage longitudinal data from Washington state from 2010-2019 to estimate the causal impact of teacher and paraeducator turnover on student outcomes. Population/Participants/Subjects: We examine the impact of staff turnover on students with and without disabilities. We assess the impact of turnover among four distinct staffing categories: general and special education paraeducators and teachers. Intervention/Program/Practice: In our analysis, we define turnover as the proportion of individuals in each school and year who either leave their school or leave the workforce the following year. Aligned with prior studies of teacher turnover (Ronfeldt et al., 2013), we use lagged turnover as our main predictor. Research Design: To estimate the causal impact of turnover on student outcomes, we use a school-and-year fixed effects model and control for a variety of time-variant characteristics of each school (e.g., student demographics and turnover among other staffing categories). Further, we interact turnover rates on student disability status. As such, we can identify the unique effects of turnover among each of our four focal staffing categories on students with and without disabilities. Given that our turnover and outcome measures are aggregated to the school level, our estimates likely reflect an underestimate of the true impact of turnover on student outcomes. Data Collection and Analysis: To calculate turnover rates by school and year, we use the S-275 dataset. The S-275 is a publicly available dataset that tracks demographic characteristics and job assignments of all individual employees in Washington state public schools. Our outcomes of interest in this analysis are test scores in ELA and mathematics among students in grades 4-8, aggregated to the school level. Finally, we use data from the National Center of Education Statistics (NCES) on urbanicity to answer our third research question regarding the relationship between school contextual differences and the impacts of staff turnover. Findings/Results: Our analysis indicates that turnover has a significant impact on student test scores, with varying effects among different position types and student groups. We find particularly notable effects of teacher turnover on ELA scores and at the elementary level. For general education students, we can predict math growth scores to be 0.06 SD lower after a year of 100% turnover among general education teachers, compared to a year of no turnover (0.09 SD in elementary school). In ELA, scores are 0.04 SD lower (0.05 SD in elementary school). For students with disabilities in ELA, we can predict that student growth scores will be 0.08 SD lower in years in which 100% of general education teachers turn over. In math, we can predict a corresponding 0.02 SD decrease. Notably, we also find statistically significant, but smaller, impacts for special education teacher turnover (-0.02 SD in math, -0.01 SD in ELA) and general education paraeducator turnover (-0.01 SD in ELA) for students with disabilities in elementary schools. Each of these estimates is significant at the 0.05 alpha level. Conclusions: Overall, our findings indicate that turnover can have a detrimental impact on general and special education student test scores, across both general and special education teachers and paraeducators. Importantly, this reflects that it is essential to consider the multitude of personnel in classrooms, especially for students with disabilities, 60% of whom are served by both general and special education teachers and who are also highly likely to be supported by paraeducators. Given high and rising turnover rates among all four personnel categories examined in this study, as well as the variety of policy interventions dedicated to abating educator turnover, our findings contribute to addressing the field's urgent need to understand the impacts of staff turnover among multiple student groups and can serve to inform educator workforce policy.