The Magnitude, Destinations, and Determinants of Mathematics and Science Teacher Turnover

Author(s): May, Henry; Ingersoll, Richard
Publication: Educational Evaluation & Policy Analysis
Published On: 8/14/2012

Few educational issues have received more attention in the past two decades than the challenge of staffing the nation's classrooms with qualified mathematics and science teachers. One of the most prominent explanations concerning this challenge focuses on teacher shortages, but there has been relatively less attention paid to the role of mathematics and science teacher turnover, migration, mobility, and attrition.

To examine the magnitude, destinations, and determinants of mathematics and science teacher turnover.


Over the past two decades, rates of mathematics and science teacher turnover have increased but, contrary to conventional wisdom, have not been consistently different than those of other teachers. Also, mathematics and science teachers were also no more likely than other teachers to take noneducation jobs, such as in technological fields or to be working for private business or industry.
  • Magnitude of mathematics and science teacher turnover:

    • The data show that from the late 1980s to 2005, annual rates of total turnover for public school mathematics and science teachers, while fluctuating from year to year, overall rose—by 34% for mathematics and by 10% for science. But the data also show that during this period mathematics and science teachers did not move from or leave their public schools at consistently different rates from other teachers.

  • Destinations of mathematics and science teacher turnover:

    • The data show that in 2004–2005 about 25,000 of those departing their schools moved to other schools and about 26,400 left classroom teaching altogether. Of those who left classroom teaching altogether, just under a third retired. Another third of leavers were job shifters who left classroom teaching but did not leave education; they took other jobs in the larger education sector. Science teachers were a bit more likely to go into nonclassroom education jobs than were nonmath/science teachers.

    • Mathematics and science teachers were not more likely than other teachers, at a statistically significant level, to leave classroom teaching to take noneducation jobs, such as in technological fields. Of those who left for noneducation jobs, mathematics and science teachers were no more likely than others to be working for private business or industry. Likewise, relatively few left to care for family members (predominantly for pregnancy and raising children) or to enroll fulltime in university or college programs.

    • Of those who moved to other schools, a large portion were cross-school transfers within the same school district. Just over half of the migrants went to teaching jobs in other districts, most within the same state. About 5% of all public school mathematics/science movers went to private schools. Math and science teachers who moved between schools were most likely to go to schools that were similar demographically. In 2004–2005, a majority of those math/science teachers who moved from high-poverty or high-minority schools migrated to schools with similar poverty and minority enrollments. However, because math/science teachers in poor, minority, and urban public schools had far higher rates of out-migration, there ended up being a net gain and loss for schools, according to school demographic differences, so the net result is a large annual asymmetric reshuffling within the school system of a significant portion (about 25,000 math/science migrants in 2004–2005) of the math/science teaching force, with a net loss on the part of poor, minority, and urban schools and a net gain to nonpoor, nonminority, and suburban schools. These patterns are similar for the nonmath/science portion of the teaching force.

  • Determinants of mathematics and science teacher turnover:

    • Individual characteristics: the age of teachers was the most salient predictor of the likelihood of their turnover. Both younger (less than 30) and older (greater than 50) teachers were more likely to depart than are middle-aged teachers. Male teachers were slightly more likely to depart than were female teachers. Minority teachers were not more or less likely to depart than were white teachers.

    • School and organizational characteristics:

      • School poverty stood out as a key variable. In general, teachers had statistically significantly higher rates of turnover in higher poverty schools than in lower poverty schools. This poverty effect was no different for mathematics or science teachers than for other subject areas.

      • After controlling for other factors, teachers in rural schools were up to 20% less likely to depart than were those in urban schools.

      • In schools with lower levels of student discipline problems, turnover rates were distinctly lower for both mathematics/science teachers and other teachers.

      • Schools that provide better principal leadership and administrative support (as reported by teachers) experienced distinctly lower turnover rates.

      • In schools where teachers reported that necessary materials, such as textbooks and supplies were available, turnover was lower for all teachers.

      • Schools with higher levels of schoolwide faculty decision-making influence had lower levels of turnover.

      • The strongest organizational predictors: for mathematics teachers the degree of individual classroom autonomy held by teachers, the provision of useful content-focused professional development, useful professional development concerning student discipline and classroom management, and the degree of student discipline problems. The strongest factors for science teachers: the maximum salary offered by school districts, the degree of student discipline problems in schools, and useful content-focused professional development.

Policy Implications/Recommendations:
  • A persistent minority of schools have continued to report serious problems filling their math and science teaching openings. The data indicate that new teacher production and recruitment strategies alone do not directly address a major root source of mathematics and science teacher staffing problems--turnover. Lowering the rate of turnover of mathematics and science teachers brought into teaching by various recruitment initiatives could prevent the loss of recruitment investments and also help to lessen the ongoing need for creating new recruitment initiatives. All this suggests the efficacy of developing teacher recruitment and retention initiatives together.

  • For math teachers, by far the strongest predictor was the degree of individual classroom autonomy held by teachers in schools in regard to content, texts, materials, techniques, and grading in their courses. For science teachers, the strongest factor was the maximum teacher salary offered by school districts and the degree of teacher classroom autonomy was not a strong factor. These findings provide support for the authors' theoretical perspective that school organization, management, and leadership matter. Schools exhibiting more characteristics associated with effective organization, and more of the indicators associated with professionalized workplaces, had significantly better retention of math and science teachers.

Research Design:
The analysis is divided into two stages: 1) descriptive statistics and 2) detailed multiple logistic regression analysis.

National Center for Education Statistics’ (NCES) nationally representative Schools and Staffing Survey (SASS) and its supplement, the Teacher Follow-Up Survey (TFS) data primarily from the 2003–2004 SASS and the 2004–2005 TFS.

Year data is from:


Data Collection and Analysis:
The analyses compares qualified mathematics teachers, with qualified science teachers, with all other teachers. The analysis is divided into two stages. In the first stage, the authors present mostly descriptive statistics to address the three research questions (turnover magnitude, destinations, and determinants). In the second stage, they follow up with a detailed multiple logistic regression analysis of the predictors of turnover to further address the determinants question.


Reference in this Web site to any specific commercial products, processes or services, or the use of any trade, firm or corporation name is for the information and convenience of the public, and does not constitute endorsement or recommendation by the Education Commission of the States. Please contact Kathy Christie at 303-299-3613 or kchristie@ecs.org for further information regarding the information posting standards and user policies of the Education Commission of the States.