Decades of research have shown that teachers are the single most important school-based factor in student achievement. The research base on teacher attrition continues to grow as teacher turnover remains a salient issue for many schools, particularly in economically disadvantaged districts. A high level of turnover is negatively associated with student achievement and there are monetary (and human capital) costs of replacing teachers. Moreover, teacher mobility patterns play an important role in the equitable education of all students, and there is strong evidence of inequities in access to highly effective instruction across schools and districts.
Taking advantage of the robust literature on factors of teacher attrition and retention, working with Lam Pham and Michael Crouch at Vanderbilt, we conducted an exhaustive search reviewing more than 25,000 scholarly records and synthesizing effects across 120 of these studies to better understand what drives teacher mobility. This post describes the findings from our analysis and offers some discussion on their policy implications.
What we know about teacher attrition and retention
For our meta-analysis, we adopt a school-centered perspective. In other words, we focus on whether teachers leave a particular school, regardless of whether they move to another school or leave the profession entirely (though our results are substantively similar when looking at attrition from the profession).
We examine factors in three primary categories that influence teacher attrition. “Personal correlates” refer to teacher characteristics—such as age, race/ethnicity, and gender—and qualifications that may shape teachers’ decisions to leave a school. “School correlates” contain factors that describe the schools and conditions in which the teachers work, including school organizational characteristics and school resources. “External correlates” are factors that come from national or state policies, such as accountability and school improvement efforts. Since our study includes several dozen factors across the three categories, we discuss a few of the more relevant findings in this brief.
Focusing first on personal correlates, in contrast to prior narrative reviews and meta-analyses, we find female teachers are no more likely to leave than male teachers, and that teachers with graduate degrees are no more likely to leave than teachers without graduate degrees. We find teaching specialty areas, such as STEM or special education, significantly increases the odds of attrition. We do not find all minority teachers are less likely to leave; only Hispanic teachers have reduced odds of attrition relative to white teachers. We also find strong evidence that teacher satisfaction plays an important role in teacher decisions to stay in teaching. Moreover, traditionally certified teachers are less likely to leave teaching than those teachers entering through alternative routes.
In terms of school correlates, we consistently find various measures of school organizational characteristics—such as student disciplinary problems, administrative support, and professional development—strongly influence teacher turnover. In terms of school resources, we find providing basic teaching materials, such as textbooks and binders, reduces odds of attrition. We find most characteristics of the student body do not seem to influence attrition (or that the influences are small).
Finally, we find that many external correlates are associated with teacher attrition and retention. Interestingly, being evaluated, even for accountability purposes, does not seem to increase teacher attrition. In fact, the odds of attrition for teachers who are assessed are somewhat smaller than those who are not.
Compelling evidence of improving school organizational characters to reduce turnover
Our meta-analytic findings have important implications for policy, practice, and future research. We find some suggestive, if limited, evidence that retention bonuses and limiting late hiring could reduce teacher turnover. Not surprising to those in the trenches, additional supports and incentives appear necessary to keep specific types of teachers in their school—namely, STEM teachers, special education teachers, and novice teachers.
We see some evidence that improving school organizational characteristics, such as reducing student disciplinary problems and improving administrative support and teacher collaborations, can reduce the risk of turnover. We are not suggesting it would be simple to change these organizational features, but the evidence warrants further exploration. While there are efforts being made in this area, we need more research and experimentation to bear on these issues.
Could evaluation and accountability improve the teacher workforce?
Despite concerns of potential negative consequences of teacher evaluation and accountability, we do not find that performance evaluations increase teacher exit. To the contrary, we find when teachers are evaluated and the results of their evaluations or measures of effectiveness are made available, teachers are no more likely to leave. In fact, the evidence suggests that teachers may be enticed to stay as they are provided with some urgency, sense of empowerment, and evidence of areas for professional improvement. This holds true even when teacher evaluations are being used for accountability and pay raises.
Related, teachers in merit-based pay programs are less likely to leave teaching than those who are not. This is noteworthy when paired with the fact that we find evaluation and accountability policies tend to be associated with keeping the most effective teachers as measured by value-added scores and removing the least effective teachers. Although differentiated compensation can be a contentious practice, it does provide suggestive evidence of the important role that strategic compensation reforms can play in improving the composition of the teacher workforce. We have a separate paper that examines merit pay in depth.
Ultimately, while there may be negative consequences and warranted concerns about teacher evaluation and accountability policies, they are positively perceived by some teachers and have more beneficial effects than previously recognized. Evaluation and accountability practices may improve the teacher workforce and reduce turnover.
Areas in need of further development and research
It is important to put findings from our meta-analysis in context of future research on teacher turnover. Two areas requiring further development and exploration are relational demography (e.g., the race-matching between teachers and principals, or between teachers and students) and school improvement. There are only a few studies on the relationship between relational demography and attrition. The same holds true for how school reforms and research-practice partnerships influence teacher attrition and retention. The nascent evidence, however, suggests relational demography, particularly parity between principal-teacher, may reduce turnover.
While there is an ever larger proportion of studies using quasi-experimental techniques to estimate causal effects of programs and policies on teacher attrition and retention, we need to further explore policy levers that can positively impact the teacher labor force and improve the educational opportunities for students from traditionally marginalized backgrounds. In an effort to propel this dialogue, our meta-analysis offers nuance to many commonly held beliefs on teacher turnover while providing new suggestive evidence of what we can do to positively impact the profession. It is now time for policy, practice, and future research to push our understanding and impact further.
The Brown Center Chalkboard launched in January 2013 as a weekly series of new analyses of policy, research, and practice relevant to U.S. education.
In July 2015, the Chalkboard was re-launched as a Brookings blog in order to offer more frequent, timely, and diverse content. Contributors to both the original paper series and current blog are committed to bringing evidence to bear on the debates around education policy in America.