Early Career STEM Teacher Retention in Texas: A Multi-Level Hazard Analysis

Background: Quality of classroom instruction is widely recognized as one of the most important school-based factors connected with student success in K12 education, and years of experience has been repeatedly identified as a key factor connected to teacher quality. However, many teachers do not remain in the profession past their first three to five years of service and thus cannot gain the experience needed to maximize student success. High levels of early career attrition among secondary teachers in science, technology, engineering, and mathematics (STEM) fields impose a variety of costs on public schools including financial costs associated with hiring and training new teachers, loss of capacity to build instructional expertise among faculty, and lack of access to high-quality teaching candidates resulting from teacher shortages. These costs are disproportionately borne by public schools which serve high need student populations. Purpose: This study examined potential determinants of attrition from public school service among secondary STEM teachers during their first five years of service within Texas public schools. The study addressed four research questions: (1) Do the individual backgrounds and experiences of early career secondary STEM teachers influence risk of attrition during the first 5 years of service within public schools in Texas? (2) Does risk of attrition vary across campuses? (3) What individual-level factors are associated with between-campus variation in risk of attrition? (4) Do organizational-level factors differentially affect risk of attrition for teachers with different individual backgrounds and experiences? Methods: Using a set of official employment and certification records for secondary STEM teachers in Texas, combined with publicly available data for public secondary schools, this study examined early career trajectories for 25,822 serving in secondary STEM teachers serving within 4,397 secondary campuses. Teachers were grouped in to six cohorts based on their first year of teaching in Texas public schools (2009-2014). A set of 66 individual and organizational-level explanatory factors related to teacher characteristics, working conditions, and broader economic conditions were included in a series of multi-level discrete-time hazard models and tested to estimate likelihood of attrition from public school service during the first five years of employment. Results: Across all cohorts, 42.6% of new teachers left public school service within five years, and teachers experienced consistent annual risks of attrition of between 10% and 12%. Factors associated with risk of attrition included certification pathways, STEM certification fields, STEM teaching assignments, number of subject area assignments, and district types. Identified levels of between-campus variation in risk of attrition were limited (> 5%), and few individual-level factors varied across campuses (e.g., lack of teaching certification, out-of-state certification). Inclusion of cross-level effects substantially reduced variation in the effect of lack of certification on risk of attrition with service in charter school districts identified as an important predictor. Conclusion: Primary contributions of this study to teacher retention research, policy, and practice included demonstrating the importance of disaggregating STEM teaching by subspecialty, confirming the value of university-based certification programs, providing evidence for the need to support beginning secondary STEM teachers during their first five years of service, and finding that limiting the number of distinct subject area teaching assignments is associated with reduced risk of attrition.