School accountability may be the most controversial and significant of all contemporary U.S. education reforms. The accountability movement began in the 1990s as some states initiated various combinations of incentives and sanctions for schools based on student test scores, under the theory that this combination of carrots and sticks would lead to improvements in school performance. In January 2002, accountability gained major impetus as a nationwide reform with the advent of the No Child Left Behind Act (NCLB), revised in 2016 as the Every Student Succeeds Act.
The impact of accountability on U.S. schools, for good or ill, is a subject of debate and research. Recently, we studied an aspect of accountability that had previously received little attention (Ingersoll, Merrill, & May, 2016). We asked, do accountability reforms affect public schools' ability to retain their teachers?
In theory, by increasing assessment and scrutiny, accountability reforms could place new pressure on teachers and hence increase teacher turnover, especially in lower-achieving schools. On the other hand, an increased focus on school accountability and performance could result in improved school leadership and management, leading to better working conditions and higher teacher retention.
Why We Need to Understand Teacher Turnover
Elementary and secondary teaching has long been marked by relatively high rates of annual turnover. By analyzing national data, we have found that attrition of teachers is similar to that of police officers, higher than nurses, and far higher than lawyers, engineers, architects, pharmacists, or academics. Moreover, the data show that the teaching force has slowly but steadily become less stable in recent years (Ingersoll, Merrill, & Stuckey 2014).
Some departure of teachers from schools is, of course, normal, inevitable, and even beneficial. Some teachers leave classroom teaching to pursue administrative positions or other education-related roles. Others leave the classroom because they discover that teaching isn't right for them. Some turnover is due to the termination of low-performing teachers. But regardless of the reason, none of these departures are cost-free. All teachers who depart leave a space behind, which takes time and effort to refill.
In earlier research, we have documented that teacher turnover is the major factor behind teacher shortages, especially in math and science (Ingersoll & Perda, 2010); and that certain working conditions and leadership actions affect teacher turnover (Ingersoll & May, 2012). The data make it clear that if we want to ensure that all students are taught by qualified teachers, many schools must pay more attention to teacher retention.
Our current study examined whether each of the typical steps involved in the implementation of accountability measures—establishing standards, using standardized assessments to measure whether a school's students meet the standards, and applying rewards or sanctions—is related to the subsequent departure of teachers from specific schools. We also explored how teachers' working conditions—the quality of school leadership, the amount of classroom resources and support provided to teachers, the level of schoolwide faculty influence over decision making, and the degree of autonomy teachers have in their classrooms—affect the relationship between school accountability and teacher turnover. We felt that the results of this study would be useful to practitioners by helping individual schools and districts establish working conditions that mitigate any potential negative effects of accountability reforms.
Our data source was the nationally representative Schools and Staffing Survey (SASS) and its supplement, the Teacher Followup Survey (TFS). Together, these are the largest and most comprehensive data sources available on elementary and secondary teachers and schools. We focused in particular on the 2003–2004 SASS and 2004–2005 TFS, conducted two and three years after the advent of No Child Left Behind. These years are a useful point at which to examine accountability because the reforms had only recently been mandated nationally; hence, we would expect large variations in design and implementation across schools. We collected data on the standards established and the performance assessed in schools in 2002–2003; rewards or sanctions subsequently applied to schools in 2003–2004; and teacher turnover between 2003–2004 and 2004–2005.
The data revealed two striking findings: (1) Some steps in school accountability were strongly related to teacher turnover, and some were not. (2) The impact of accountability strongly depended on teachers' working conditions in their school.
The Growth and Impact of Accountability
Following the passage of NCLB, performance-based school accountability quickly became widespread. Almost all public schools in the United States have been subject to performance standards and assessments established by their district or state. Our analysis of the data found that, of those schools that were assessed in the 2002–2003 school year, more than half (54 percent) passed all the performance standards, and 39 percent of these schools subsequently received some kind of reward. Of the 46 percent of evaluated schools that failed to pass some or all of the standards in 2002–2003, just over half were subsequently subjected to some kind of penalties or sanctions.
Sanctions varied in their degree of specificity, seriousness, and frequency. The most common result of not meeting some or all of the standards was a requirement to write a school improvement plan. Next most frequent was a tighter sanction—being put on an evaluation cycle with required improvements by specific dates. Far less common were more punitive consequences, such as a reduction in resources, school takeover, or reconstitution of administrative and teaching staff.
The data also show that these reforms had an impact on teachers and their teaching practices by 2003–2004, only one year into No Child Left Behind (see fig. 1). Ninety-one percent of teachers reported that their students participated that year in required state or district assessments. And, regardless of whether a teacher's students were tested in that teacher's subject, the use of standards and standardized test scores had become a ubiquitous part of life for the great majority of teachers. Most teachers had access to their students' scores on state or district achievement tests. Most reported that they used state or district standards to guide their teaching and used test results to adjust their classroom curriculum and to assess weaknesses in their own content knowledge or teaching practices. Moreover, this impact appears to have had some bite. Almost one-third of all public teachers reported they were "somewhat" or "strongly" worried about their job security because of the performance of their students on state or local tests.
Figure 1. Percent of Public School Teachers Who Reported They Experienced the Effects of Accountability Policies, School Year 2003–2004
Source: What are the Effects of School Accountability on Teacher Turnover? (CPRE Research Report), by R. Ingersoll, L. Merrill, & H. May, 2016. Philadelphia: Consortium for Policy Research in Education, University of Pennsylvania. Copyright © 2016 by Richard Ingersoll. Adapted with permission.
From the perspective of teacher job satisfaction, this impact has not always been viewed as benign. Indeed, less than half (45 percent) of responding teachers reported that state or district standards had a positive influence on their satisfaction with teaching. Our next question: What impact did these reforms have on teacher retention and turnover?
How Accountability Affects Teacher Turnover
After controlling for the background characteristics of teachers and schools, our statistical analyses of the SASS and TFS data showed that some steps in school accountability were related to teacher turnover, and some were not. Perhaps surprisingly, having performance standards and state or district assessments in a school did not have a negative effect on teacher retention—these factors did not, in and of themselves, drive out teachers.
However, how schools performed on assessments did affect retention and turnover. Not surprisingly, successful schools had better retention, and less-successful schools had worse retention. Interestingly, rewards given to higher-performing schools did little to improve these schools' already-higher retention. In contrast, sanctions applied to lower-performing schools did a lot to worsen their already-lower retention. One of the most consequential sanctions for teacher turnover was a school being put on an evaluation cycle with specific deadlines for improvement. Thirty percent of the low-performing schools were subject to this sanction, either alone or in combination with other sanctions. These schools had significantly higher turnover than did low-performing schools that were not subject to this sanction.
Is there anything that low-performing schools, especially those subject to sanctions, can do to ameliorate their losses of teachers? Is it possible to implement accountability in a way that does not exacerbate low-performing schools' problems by driving out more teachers?
The data show large school-to-school differences in the four working conditions we examined: the quality of school leadership, the amount of classroom resources and support provided to teachers, the level of schoolwide faculty influence over decision making, and the degree of autonomy teachers have in their classrooms. And these differences in working conditions mattered for retention. Teachers in schools with higher levels of leadership support, classroom resources, schoolwide influence, or classroom autonomy all had significantly lower turnover, after controlling for the background characteristics of the teachers and schools as well as school performance, rewards, or sanctions.
One of our four working conditions—classroom teacher autonomy—was especially powerful in ameliorating the effects of accountability in low-performing schools. The relationship of sanctions to teacher turnover in these schools strongly depended on how much autonomy teachers were allowed in their own classrooms over key issues: selecting textbooks and other instructional materials; choosing content, topics, and skills to be taught; evaluating and grading students; selecting teaching techniques; determining the amount of homework to be assigned; and disciplining students.
In general, sanctioned low-performing schools gave teachers less classroom autonomy than other schools did. But those sanctioned low-performing schools that did provide greater classroom autonomy to teachers had far lower teacher turnover. In other words, low-performing schools with sanctions had far higher turnover if their teachers were allowed less classroom autonomy, and they had far lower turnover if their teachers were allowed more autonomy.
Figure 2 illustrates the differences in the levels of teacher turnover, according to the performance of schools. On average, 12.6 percent of the teachers in the higher-performing schools departed between the 2004 and 2005 school years. Turnover was higher (15.7 percent) in low-performing schools (those that failed to pass some or all of the performance standards). Turnover was highest (20 percent) in low-performing schools that were subject to some kind of penalties or sanctions. However, among low-performing sanctioned schools, those that gave teachers greater classroom autonomy had significantly lower turnover—in fact, the rate (12.2 percent) was similar to that in higher-performing schools. Thus, our analysis shows that it is not inevitable that sanctioned low-performing schools lose more teachers.
Figure 2. Percent of Teachers Who Left Between 2004 and 2005, by School Performance, Sanctions, and Level of Teacher Autonomy: School Year 2004–2005
Source: What are the Effects of School Accountability on Teacher Turnover? (CPRE Research Report), by R. Ingersoll, L. Merrill, & H. May, 2016. Philadelphia: Consortium for Policy Research in Education, University of Pennsylvania. Copyright © 2016 by Richard Ingersoll. Adapted with permission.
Give Teachers the Tools They Need
Nothing in our study suggests that school accountability is a bad thing. Proponents of accountability reforms have identified important issues and problems; establishing standards and requiring transparency about school performance are reasonable and necessary actions. The public has a right and, indeed, an obligation to be concerned with the performance of schools and teachers. Moreover, there is no question that some schools and teachers are performing poorly, in one way or another.
However, the evidence shows that implementation of accountability reforms can contribute to school performance problems; if we overlook that fact, such reforms may backfire. If the way schools are managed and organized undermines the ability of teachers to feel successful in helping students learn—the very reason many of them went into teaching in the first place—such reforms may not only fail to solve the problems they seek to address, but may also end up making things worse.
In plain terms, it stands to reason that if teachers are to successfully meet standards, schools must be organized in ways that give teachers the tools, capabilities, and resources they need to do so. Our data suggest that a key resource that teachers value is having sufficient control over instructional decisions in their classrooms.
This finding is especially relevant for hard-to-staff and low-performing schools (often the same entities). In addition to school-level accountability, federal and state regulations also mandate that schools successfully staff all of their core academic classrooms with highly qualified teachers. If low-performing schools are to ensure that all their classrooms are staffed with qualified teachers, they must give their staff sufficient control and autonomy in their classrooms.
Authors' Note: This research was supported by a grant (#0814295) from the Research and Evaluation on Education in Science and Engineering (REESE) program of the National Science Foundation. Lisa Merrill's work was also supported, in part, by the Institute of Education Sciences, U.S. Department of Education, through Grant #R305B090015 to the University of Pennsylvania. Opinions in this paper reflect those of the authors and do not necessarily reflect those of the granting agencies.
