Spillover Effects of Recruiting Teachers for School Turnaround: Evidence From Tennessee

Abstract

Many districts and states have implemented incentives to recruit teachers to low-performing schools, and previous research has found evidence that these incentives are effective at attracting teachers. However, effects on the schools and students these teachers leave behind have not been examined. This study focuses on the spillover effects of recruiting effective teachers to Tennessee’s Innovation Zone (iZone) schools. We find the short-term effects of losing these teachers range from −0.04 to −0.12 SDs in student test score gains, with larger negative effects when more effective teachers leave. However, combining both these negative effects in schools teachers leave and the positive effects in iZone schools yields overall net positive effects.

Keywords

achievement at-risk students educational policy educational reform equity school/teacher effectiveness econometric analysis longitudinal studies policy analysis quasi-experimental analysis regression analyses secondary data analysis

School reform efforts nationwide operate under multifaceted theories of action, but recruiting effective teachers into low-performing schools remains one of the most prevalent improvement strategies across myriad reform models (Clotfelter et al., 2008; Cowan & Goldhaber, 2015; Glazerman et al., 2013; Springer et al., 2016; Steele et al., 2010). As efforts to recruit teachers into high-need schools have been pervasive, these policies have been evaluated extensively with evidence of positive effects on student achievement (Glazerman et al., 2013; Papay & Hannon, 2018; Swain et al., 2019; Zimmer et al., 2017). However, while a fair amount of research has examined the effects of teacher recruitment on schools that receive them (i.e., receiving schools), studies have largely ignored spillover effects on schools that lose these teachers (i.e., sending schools). In fact, evaluations of teacher recruitment policies tend to make policy conclusions based solely on effects in receiving schools without considering the broader implications for sending schools, despite robust evidence that teacher turnover can adversely affect student achievement (Hanushek et al., 2016; Henry & Redding, 2020; Ronfeldt et al., 2013).

To more fully account for the broader effects of teacher recruitment efforts, we focus on a school reform policy in Tennessee known as the Innovation Zones (iZones). We describe the iZone reform model in detail below, but prior research has found that iZone schools successfully recruited effective teachers (Henry et al., 2017) and increased student achievement (Pham et al., 2020; Zimmer et al., 2017). Moreover, attracting effective teachers was a mechanism that mediated the positive iZone effects (Henry et al., 2020). In this article, we estimate the spillover effects on sending schools when they lose teachers to the iZones. Specifically, we ask: To what extent did iZone schools’ recruitment efforts affect student achievement in sending schools? To answer this question, we use value-added measures of student performance in a series of fixed-effects models that follow recent research on the effects of teacher turnover (Henry & Redding, 2020; Ronfeldt et al., 2013). These models allow us to examine changes in student test score gains in sending schools after teachers leave for an iZone school.

By highlighting sending schools, we make several contributions to the literature on teacher recruitment in the school reform context. First, given the dearth of information on sending schools, we first examine their characteristics to better illuminate the types of schools most likely to lose teachers. If sending schools are generally high-performing and well-resourced, they may have natural advantages in recruiting new teachers or have the ability to counteract negative effects from losing teachers. However, if sending schools are themselves low-performing, losing teachers could further destabilize them, suggesting a zero-sum situation where sending schools’ performance must decline to support improvement in receiving schools, at least in the short term.¹ Second, we examine the characteristics of teachers who leave sending schools. Specifically, we compare teachers who (a) leave for iZone schools with (b) those who leave for non-iZone schools and (c) those who stay in sending schools. By doing so, we contribute a more nuanced understanding of the kinds of teachers that sending schools lose in response to teacher recruitment efforts. Third, we broaden the narrative on teacher recruitment policies by estimating the effects on sending schools rather than focusing exclusively on receiving schools. Using these effect estimates in sending schools, we then calculate the net effects of iZone teacher recruitment policies, accounting for the effect in both iZone schools and sending schools. Calculating net effects allows us to demonstrate how evaluations of teacher recruitment policies can account for both direct effects on receiving schools and spillover effects in sending schools. To our knowledge, this is the first study to consider both teacher recruitment efforts on sending schools and net effects.

Our findings suggest that teacher transfers from sending schools increased after the iZones were established and that students in sending schools perform worse when entering grades and subjects previously taught by a teacher who left for an iZone school, especially if the teacher who left was an effective teacher. We also find that most sending schools are located in the same district and, on average, are only slightly higher-performing than iZone schools. These findings highlight an important and often ignored spillover effect of teacher recruitment policies: significant negative impacts on nearby sending schools that are themselves low-performing. However, when accounting for both negative effects in sending schools and positive effects in iZone schools, we find that gains in iZone schools outweighed the losses in sending schools. That is, the overall effect of iZone teacher recruitment incentives is positive. These results have direct implications for the short-run unintended consequences of teacher recruitment policies and highlight the importance of accounting for both direct and spillover effects when evaluating teacher recruitment efforts.

Literature Review

The existing research consistently finds that teacher quality matters. Students taught by more effective teachers have higher test score gains, more positive non-cognitive outcomes, and better long-term outcomes (Aaronson et al., 2007; Chetty et al., 2014; Hanushek, 2011; Jackson, 2018; Jackson et al., 2014; Koedel & Betts, 2007; Rivkin et al., 2005; Rockoff, 2004; Sanders et al., 1997). Unfortunately, research also finds that schools with high-poverty, high-minority, and low-performing students employ less effective and less experienced teachers (Glazerman & Max, 2011; Goldhaber et al., 2015; Isenberg et al., 2013; Loeb et al., 2012; Sass et al., 2012; Steele et al., 2015). To disrupt this unequal distribution of teachers, educational leaders and policymakers have implemented a number of policies to recruit effective teachers into low-performing schools, including financial incentives and even involuntary transfers (Grissom et al., 2014; Springer et al., 2016). A growing body of research provides credible evidence that these recruitment efforts have increased the number of effective teachers in high-poverty, high-minority, and low-performing schools, although findings about the retention of these teachers are mixed. Two studies (Cowan & Goldhaber, 2015; Steele et al., 2010) showed that financial incentives can attract more effective teachers into low-performing schools, but these bonuses did not affect retention of those teachers in the respective schools. In contrast, research by Clotfelter and colleagues (2008) and Springer et al. (2016) show that retention bonuses for effective teachers in low-performing schools have had positive effects on teacher retention, and a follow-up study by Swain et al. (2019) found that the positive effects also translated into increases in student achievement. Furthermore, a large randomized experiment that provided bonuses to attract effective teachers into low-performing schools in 10 large school districts across seven states found positive effects on teacher recruitment, teacher retention, and student achievement (Glazerman et al., 2013). These prior studies have identified a plausible causal effect of teacher recruitment incentives in the receiving school, but no studies have addressed spillover effects that may result when teachers leave their sending schools as a result of such recruitment policies.

Teacher recruitment efforts could cause negative spillover effects on students by increasing turnover in sending schools. Researchers have posited that increasing teacher turnover can negatively affect student achievement through two main pathways—“compositional” effects and “disruptive” effects (Ronfeldt et al., 2013). “Compositional” effects occur when there is a difference in quality between the teachers who leave and the teachers who replace them. If exiting teachers are replaced by less effective counterparts, student achievement will likely decrease. If exiting teachers are replaced by more effective counterparts, student achievement will likely increase (Ronfeldt et al., 2013). Although prior studies have not specifically examined the compositional effect directly, research finds that less effective teachers are more likely to leave than effective teachers (Goldhaber et al., 2011; Hanushek & Rivkin, 2010; Murnane, 1984), suggesting a positive compositional effect on student achievement. A more recent descriptive study finds that incoming teachers are more effective than or just as effective as exiting teachers in Tennessee (Henry et al., 2017), again suggesting a potentially positive compositional effect of typical teacher turnover on student achievement. However, without accounting for the effectiveness of incoming teachers, these prior studies cannot definitively assess the overall compositional effect.

Regardless of exiting and replacement teacher quality, turnover can still have “disruptive” effects that extend beyond the quality of these two teachers. During their tenure, teachers build relationships with various stakeholders at the school. They form a community of learning and develop a level of trust with students, parents, and other teachers, among others in the building. When teachers then leave a school, those relationships are disrupted and can create a negative impact on school climate and interactions among colleagues, which can then affect student learning and achievement (Bryk & Schneider, 2002; Guin, 2004; Ronfeldt et al., 2013). Furthermore, the teacher’s organizational knowledge and professional learning leave with them, disrupting coherent implementation of instructional programs (Guin, 2004). Faculty remaining in the school must then take on additional responsibilities the leaving teacher previously served and bring new teachers up to date with instructional programs and dynamics among students, thereby taking time away from their previous responsibilities and slowing institutional progress (Ronfeldt et al., 2013).

Including both compositional and disruptive effects, prior studies have empirically examined the overall impact of teacher turnover, concluding a negative effect on student achievement. Students’ test score gains are reduced by 4% to 12% of a standard deviation (calibrated to 100% of teachers leaving; Hanushek et al., 2016; Henry & Redding, 2020; Ronfeldt et al., 2013). Although existing research has examined the effects of general teacher turnover on student achievement, no studies have explicitly evaluated the effects of turnover induced by teacher recruitment policies that intentionally draw teachers away from other schools. We address this gap in the literature by investigating spillover effects of teacher recruitment policies on sending schools. Based on prior literature, we hypothesize a negative impact on student achievement but perhaps even greater in magnitude than prior studies. Although previous studies examined turnover effects broadly and found that less effective teachers were more likely to leave, we focus on a policy that aimed to attract and draw away more effective teachers. Thus, while negative disruptive effects are expected, it is possible that compositional effects may also be negative, further contributing to the harmful effects on student achievement.

The Tennessee Context

In 2011, Tennessee received a waiver from the No Child Left Behind goal of having 100% of students proficient in reading and math by 2014. As a part of that waiver, Tennessee identified its lowest-performing 5% of schools, called Priority schools. These schools reside primarily in Tennessee’s largest cities—69 in Memphis, 6 in Nashville, 6 in Chattanooga, and 2 in two smaller school districts. In addition, the state decided that some Priority schools would undergo one of two improvement interventions including placement into iZones. Local iZones are networks of low-performing schools that are part of their local districts but are managed separately. Schools placed into an iZone are required to hire a new principal and are given increased autonomy over daily operations. The three cities housing most of Tennessee’s Priority schools all opened iZones, and most Priority schools became part of an iZone.² In the first year of Priority status, 2012–2013, Nashville opened its iZone with four schools, while Memphis opened its iZone with seven schools. The following year, Memphis added six more schools to its iZone, and Chattanooga began an iZone with five Priority schools. Finally, in 2014–2015, Memphis added four schools to its iZone.

All three districts’ iZones were tasked with improving persistently low-performing schools that had come under their management. To accomplish this goal, the state provided some parameters around strategies iZones should use. For instance, all districts had to extend learning time for students and hire new leadership to run iZone schools (for more on leadership, please see Dixon et al., 2021). Of particular interest in this study, iZones had to utilize one of the state’s key improvement strategies for its low-performing schools—recruiting and retaining high-quality teachers (United States Department of Education, 2009, 2010). To assist with teacher recruitment, the Tennessee Department of Education (TDOE) offered pay bonuses, where teachers with above average evaluation scores were eligible for a US$7,000 signing bonus in return for transferring into and staying in a Priority school for at least 2 years (Springer et al., 2016; TDOE, 2013). Individual districts, however, had the flexibility to also provide their own incentives. Memphis offered up to US$1,500 in bonuses to teachers who agreed to teach for at least 3 years in iZone schools with an additional US$1,000 annual award if teachers met district benchmarks (Burnette, 2017; Sullivan, 2013), and Hamilton provided a differentiated pay scale for iZone teachers with performance bonuses built in (Hardy, 2013). Districts’ approaches to hiring new teachers and principals also differed. For instance, Chattanooga specifically hired principals from outside the district, whereas Memphis primarily hired teachers who had prior experience teaching or leading in its community.³

Although the different iZones and even individual schools within each iZone may have implemented different strategies with differences in the resulting educational infrastructures, the iZones were required to follow state and federal guidelines which resulted in implementing similar personnel policies and financial incentives for recruiting and replacing teachers. Descriptive analyses of teacher quality and mobility in iZone schools showed an increase in the hiring of highly effective teachers overall, as determined by value-added measures and teaching experience (Zimmer et al., 2017), and after 3 years of reforms, iZone schools yielded average student test score gains that were 0.10 to 0.20 SDs higher than student achievement in other Priority schools (Zimmer et al., 2017). Furthermore, Henry and colleagues (2020) found that recruiting high-quality teachers partially mediated the positive effects of iZone reforms on student achievement. In this study, however, instead of focusing on iZone schools, we investigate whether the positive iZone results came at the expense of sending schools.

Method

Data

We use statewide administrative data on students, teachers, and schools spanning 2011–2012 to 2014–2015, provided by TDOE and managed by the Tennessee Education Research Alliance. The student data include demographic data, school assignments, and test scores on statewide end-of-grade (Grades 3 through 8) exams in reading, math, and science, and end-of-course exams (in select high school courses) also in these three subjects, which we standardize by year, grade, and subject to use as the outcome of interest. The teacher data include race/ethnicity, education level, years of experience, and Teacher Value-Added Assessment System (TVAAS) scores, an annual rating of teacher impact on students’ academic progress. TVAAS scores are based on state standardized assessment student-level growth scores and range from 1 (least effective) to 5 (most effective), with 3 and above indicating an effective teacher. Only teachers teaching tested grades and courses receive a TVAAS score.

The teacher data also allow us to track teachers’ school assignments in every year, including grade and subject assignments. Therefore, we can create a continuous variable capturing the proportion of teachers exiting grade g in school s in year t − 1 to enter an iZone school. This grade-level teacher turnover proportion serves as the main independent variable. One limitation for this operationalization of teacher turnover is that our data do not identify reasons why teachers leave their schools. Therefore, it is difficult to ascertain whether teachers who left for the iZones would have left regardless of the iZone recruitment efforts. However, we can examine turnover trends for the sending schools (see Supplementary Appendix Figure A1 in the online version of the journal) and use an interrupted time series model (see Supplementary Appendix Table A1 in the online version of the journal) to show that the number of teachers transferring from sending schools to receiving schools increase significantly when the receiving school joins an iZone, providing evidence that teachers did indeed transfer into iZone schools because of iZone recruitment efforts. Nonetheless, as described further below, we also control for other teacher turnover to help separate the effect of losing teachers to the iZone from teacher turnover that would have occurred in the absence of iZone recruitment efforts.⁴

Empirical Framework and Samples

We use a series of value-added equations to estimate student achievement gains along with student, school, year, and grade fixed effects, similar to methods used in recent studies of teacher turnover (Henry & Redding, 2020; Ronfeldt et al., 2013). In particular, we are concerned with endogeneity from pre-existing school characteristics, because sending schools likely differ from schools that did not lose teachers to the iZones. Therefore, our preferred model utilizes a school-by-year fixed effect, which leverages variation in teacher turnover across grades within the same school and year. The school-by-year fixed effect allows us to compare students in one grade with other students in a different grade in the same school and year, where teacher turnover rates vary from grade to grade. This approach controls for observed and unobserved school-by-year characteristics that affect both teacher turnover and student achievement. For instance, principal turnover or effectiveness at a school in one particular year may simultaneously influence student achievement and teacher turnover, biasing estimates of how turnover affects achievement. With the school-by-year fixed effect, we control for omitted confounders specific to the school and year, such as principal turnover and effectiveness. The model assumes that the effects of teacher turnover on student achievement are comparable across grade levels. We model this approach using

\begin{array}{l} y_{i g s t} = β_{0} + β_{1} i Z o n e T c h r G r a d e T u r n o v e r_{g s t - 1} \\ + β_{2} O t h e r T c h r G r a d e T u r n o v e r_{g s t - 1} \\ + β_{3} y_{i g s t - 1} + S_{i g s t} B_{j} + γ_{s t} + e_{i g s t} \end{array}

(1)

where y represents the test score for student i in grade g, school s, and year t. iZoneTchrGradeTurnover is the proportion of teachers who left grade g in school s in year t − 1 (the year prior to a student entering the respective grade) to teach at an iZone school. β₁ is the coefficient of interest and can be interpreted as the effect of losing all teachers in the grade to the iZones. In addition to teachers leaving for iZone schools, some teachers depart sending schools for other reasons. To avoid misattributing the effects of other turnover to iZone recruitment efforts, we include a turnover control variable that captures all teacher moves except moves to an iZone school. That is, OtherTchrGradeTurnover_gst−₁ is the proportion of teachers who left grade g in school s in year t − 1 for a receiving school that is not an iZone school. This variable ensures that the effects of grade-level turnover for reasons other than leaving to join an iZone are not erroneously attributed to the effects of leaving for iZone schools. y_igst₋₁ represents the student’s test score in the year prior⁵; S_igst represents a vector of student characteristics—gender, race, economically disadvantaged status, special education status, English language learner status, and mobility status, measured as whether a student was new to the school in year t from a nonstructural move (moves that are not mandated by a school’s terminal grade); γ_st represents the school-by-year fixed effect; and e_igst is an idiosyncratic error term. Standard errors are clustered at the school level. We estimate this model separately for reading, math, and science scores, and only include teachers who teach the respective subjects in the teacher turnover variables. Equation 1 accounts for time-varying school characteristics by comparing teacher turnover to iZone schools that occurred in one grade to other grades in the same school and year that did not lose teachers to the iZone, adjusted for any other teacher turnover in the same grade.

Equation 1, however, could be biased by within-school grade-level confounders. For example, a teacher may leave her school because of ineffective peers teachings in the same grade. Ineffective grade-level peer teachers is a potential confounder that could influence both teacher turnover and student achievement but is not controlled by the school-by-year fixed effect. Therefore, we estimate a second model, replacing the school-by-year fixed effect with a school-by-grade fixed effect. The school-by-grade fixed effect model leverages variation in teacher turnover across years within the same school and grade. That is, the school-by-grade fixed-effects model compares students in one grade, school, and year with students attending the same grade and school but in different years. This approach allows us to control for omitted variables specific to the grade and school (e.g., grade-level peer effectiveness) and exploits the variance in turnover over time within the same grade and school, assuming that the effects of teacher turnover in the same grade and school are comparable across years. We model this approach using

\begin{array}{l} y_{i g s t} = β_{0} + β_{1} i Z o n e T c h r G r a d e T u r n o v e r_{g s t - 1} \\ + β_{2} O t h e r T c h r G r a d e T u r n o v e r_{g s t - 1} \\ + β_{3} y_{i g s t - 1} + S_{i g s t} B_{j} + X_{s t} B_{k} \\ + δ_{g s} + θ_{t} + e_{i g s t} \end{array}

(2)

In this specification, the within-school differences in student achievement gains before and after teachers transferred to the iZones are used to estimate the effects from losing a teacher to the iZones. In addition to student characteristics, we control for school-level characteristics—percentage of students eligible for free or reduced-price meals, percentage of students who are non-White, and percentage of mobile students (students who are new to a school due to a nonstructural move) for school s at time t—which is represented by X_st, and employ a year fixed effect θ_t to adjust for overall yearly differences. δ_gs represents the school-by-grade fixed effect. Standard errors are clustered at the school level. Again, we run this model separately for reading, math, and science scores and only include teachers who teach the respective subjects in each model.

By using both approaches, we can examine whether our results are robust to the assumptions of each model. However, the school-by-year fixed-effects approach (Equation 1) is our preferred model because we believe the factors simultaneously affecting both student achievement and teacher transfers are more likely to occur in a school from year to year rather than between grades in a single year.⁶

Finally, we extend these analyses by investigating the characteristics of teachers and sending schools that may moderate the effects of teacher turnover. First, previous literature has found that turnover among effective teachers is more harmful to student achievement, and turnover of ineffective teachers may actually improve student achievement (Adnot et al., 2017). Second, teacher turnover is typically more harmful for students in schools with more economically disadvantaged and low-performing students (Hanushek et al., 2016; Ronfeldt et al., 2013). Therefore, we examine heterogeneous effects for effective versus ineffective teachers, the upper and lower quartiles of schools’ percentage of students who are economically disadvantaged relative to the middle half of the distribution, and sending schools who were also labeled Priority (the state’s lowest performing 5% of schools) versus those that were not.

Results

Characteristics of Teachers Recruited into iZone Schools

From 2012–2013 through 2014–2015, the first 3 years of iZone operations, 653 teachers transferred into one of 26 iZone schools in Memphis, Nashville, or Chattanooga, 238 of whom taught a tested subject or grade. Of these transferring teachers, 92% moved from other schools in the same district as the receiving iZone school, 4% came from nearby or bordering districts, and 3% moved from other districts throughout the state. Table 1 shows the characteristics of teachers in sending schools disaggregated into teachers who transferred to iZone schools, teachers who transferred to other schools during the same period, and teachers who stayed in these schools. For reference, we also show these teacher characteristics for iZone teachers (prior to becoming iZone schools), teachers in Memphis, Nashville, and Chattanooga (where most of the sending schools and all iZone schools are located), and all teachers in the state. Note that we only show characteristics for teachers with TVAAS scores. However, teachers without TVAAS scores are similar in other characteristics to those with TVAAS scores. For instance, inclusive of those without TVAAS scores, teachers transferring to iZone schools were 28% White and 70% Black; 69% had master’s degrees or higher; and the average teacher had 9.0 years of experience.

Table 1

Baseline Teacher Characteristics of Sending Schools, iZone Schools, Districts With iZones, and All Tennessee Schools

Teacher characteristics	Sending schools			iZone schools	Memphis, Nashville, and Chattanooga	Tennessee
Teacher characteristics	Transferred to iZone school	Transferred to non-iZone school	Stayed in sending school	iZone schools	Memphis, Nashville, and Chattanooga	Tennessee
Race
White	27%	46%	53%	25%	59%	87%
Black	70%	52%	45%	74%	40%	12%
Years of experience	7.9	8.1	9.4	11.3	10.3	11.3
Master’s degree +	71%	56%	62%	68%	61%	57%
Effective (prior year)	60%	54%	67%	49%	67%	72%
Total teachers	238	482	6,608	394	17,171	83,903

Note. Effectiveness is determined by teachers’ Teacher Value-Added Assessment System (TVAAS) scores, which are based on student growth scores on state standardized assessments. Only teachers with TVAAS scores were included.

Table 1 shows that teachers transferring to iZone schools were primarily Black (70%, similar to those in iZone schools), had fewer years of experience (7.9), on average, and were more likely to have a master’s degree (71%), relative to teachers transferring from the same sending schools to non-iZone schools (52% Black, 8.1 years of experience, 56% with master’s degree) and those staying in sending schools (45% Black, 9.4 years of experience, 62% with master’s degree). Furthermore, 60% of teachers transferring to iZone schools earned an effective rating based on TVAAS scores in the prior year. This is greater than those transferring to non-iZone schools (54%), but lower than those staying in sending schools (67%). Relative to the overall characteristics of iZone receiving schools, teachers who transferred in were approximately the same racial/ethnic distribution, had fewer years of experience, had similar percentages of master’s degrees, and were more effective. Relative to the districts in which iZones were located and the state as a whole, teachers transferring into iZone schools were much more likely to be Black and to have a master’s degree, had fewer years of experience, and were less likely to be effective.

Characteristics of Sending Schools

The 238 teachers moving into iZone schools came from 143 different schools, averaging a loss of 1.6 teachers per school, although several schools lost as many as six teachers in 1 year and one up to 14 teachers over the 3-year period. In Table 2, we compare the baseline school-level characteristics of sending schools and iZone schools. We also include as reference the overall characteristics of all Memphis, Nashville, and Chattanooga schools and all Tennessee schools. For the most part, iZone schools served elementary and middle grades. Therefore, most sending schools were also elementary and middle schools. Sending schools had smaller percentages of minority and economically disadvantaged students than iZone schools—83% minority and 79% economically disadvantaged in sending schools compared with 97% minority and 92% economically disadvantaged in iZone schools—but greater than the average school in districts with iZones (74% minority and 71% economically disadvantaged) and much greater than the average Tennessee school (33% minority and 60% economically disadvantaged). Similarly, the sending schools, which scored 0.43 to 0.69 SDs below the state average depending on subject, were higher performing on the state’s standardized assessments than iZone schools, which scored 0.86 to 1.13 SDs below average, but worse than the average school in Memphis, Nashville, or Chattanooga (0.34–0.47 SDs below average), and much worse than the average Tennessee school (0.01 SDs) in the baseline year. In addition, approximately one quarter of sending schools were also Priority schools (the state’s lowest-performing 5% of schools), meaning many already low-performing schools lost some of their talent when the iZones began recruiting teachers. Overall, teachers that transferred to the iZone left sending schools that were slightly less disadvantaged and slightly higher-performing than the iZone schools themselves.

Table 2

Baseline School-Level Characteristics of Sending Schools, iZone Schools, Districts With iZones, and All Tennessee Schools

School characteristics	Sending schools	iZone schools	Memphis, Nashville, and Chattanooga	Tennessee
School level
Elementary	35%	50%	55%	59%
Middle	45%	35%	24%	24%
High	20%	15%	21%	17%
Percent minority	83%	97%	74%	33%
Percent economically disadvantaged	79%	92%	71%	60%
Priority schools (lowest-performing)	27%	100%	17%	5%
Reading score^a	−0.58	−1.03	−0.35	0.01
Math score^a	−0.43	−0.86	−0.34	0.01
Science score^a	−0.69	−1.13	−0.47	0.01
Total schools	142	26	478	1,615

Alternative schools are excluded.

Test scores represent average standardized test scores in years prior to teacher recruitment/loss and are standardized at the state level.

Spillover Effects

In Table 3, we display the estimated effects of teachers leaving for the iZone on student test scores of the grades and subjects in the sending schools in the year after teachers leave. Columns 1, 3, and 5 provide the results of our preferred model—the school-by-year fixed-effects model; Columns 2, 4, and 6 provide the results of the school-by-grade fixed-effects model for reading, math, and science, respectively.⁷ For each model, we also display the coefficient for other teacher turnover as a comparison and indicate in bold coefficient estimates that are statistically different from the coefficient estimates for teacher turnover to the iZone.⁸ Each of the coefficient estimates should be interpreted as the change in test score gains for students entering a grade in which all teachers left the previous year. On average, grades that lost reading, math, and science teachers to the iZone lost 53%, 62%, and 65% of their grade-level teachers to iZone schools, respectively.⁹ Therefore, to correctly calibrate the actual effect of teacher turnover to iZone schools, the estimates should be multiplied by 0.53, 0.62, and 0.65 for reading, math, and science, respectively.

Table 3

Estimates of the Effects of Teacher Turnover to iZone Schools on Student Achievement

Variable	Reading		Math		Science
Variable	(1)	(2)	(3)	(4)	(5)	(6)
Teacher turnover to iZone	−0.10**	−0.07^†	−0.07	−0.14*	−0.14*	−0.19**
	(0.04)	(0.04)	(0.09)	(0.07)	(0.05)	(0.06)
	[−0.053]	[−0.037]	[−0.043]	[−0.087]	[−0.091]	[−0.124]
Other teacher turnover	−0.02	−0.01	−0.13	−0.14	0.04	−0.05
	(0.05)	(0.05)	(0.11)	(0.12)	(0.10)	(0.09)
School × Year FE	X		X		X
School × Grade FE		X		X		X
R ²	0.63	0.62	0.48	0.50	0.55	0.53
N	17,419	21,293	17,048	20,415	14,414	18,193

Note. Standard errors are clustered by school and indicated in parentheses. FE = fixed effect; FRPL = free or reduced-price meals; ELL = English Language Learner.

The main coefficient estimates can be interpreted as the change in test score gains for students entering a grade in which 100% of teachers left the previous year. On average, grades that lost reading, math, and science teachers to the iZone lost 53%, 62%, and 65% of their grade-level teachers to iZones, respectively. Therefore, to correctly calibrate the actual effect of teacher turnover to iZone schools, the main coefficient estimates should be multiplied by 0.53, 0.62, and 0.65 for reading, math, and science, respectively. We have included the result of these products in brackets.

Student Controls: Gender, Race, FRPL status, Special Education status, ELL status, Mobility Status, Prior Reading Test Score, Prior Math Test Score, Prior Science Test Score.

School Controls: Percent Minority, Percent FRPL, Percent Student Mobility, School Level.

Estimates for Other teacher turnover in bold indicate statistically significant differences from estimates for Teacher turnover to iZone in the respective model.

†

p < .10. * p < .05. ** p < .01. ***p < .001.

Our preferred model shows that students in grades that lost all of their reading teachers to the iZone scored 0.10 SDs lower on their reading assessment than students in the same school and year that did not lose any teachers to the iZones. As grades that lost reading teachers to the iZone lost 53% of their grade-level teachers on average, the observed average effect estimate is −0.053 SDs in reading (−0.10 × 0.53; displayed in brackets in Table 3). In the school-by-grade model, our coefficient of interest is statistically significant at the 10% alpha level—students in grades that lost 100% of their reading teachers to the iZones scored 0.07 SDs lower on their reading assessments, which translates to approximately 0.037 SDs considering the average percentage of teachers transferring to iZone schools. In neither model did we find any positive or negative effects of losing reading teachers who moved to non-iZone schools.

In math, we find null effects in our preferred model; however, in the school-by-grade fixed-effects model, our estimate suggests that students in grades losing all math teachers to the iZone would score 0.14 SDs lower than they would if none of their grade-level math teachers left for the iZone. Taking into account the average percent of grade-level teachers lost, 62%, this translates to a 0.087 SD loss. The effect of losing teachers to other reasons is not statistically distinguishable from zero.

In science, both models yield statistically significant negative effects that are the largest of all three subjects. Students entering grades that lost all of their science teachers to the iZone scored 0.14 to 0.19 SDs lower than the respective comparison groups, which translates to about 0.091 to 0.124 SDs after accounting for the percentage of teachers who transferred to iZone schools. The effect of losing a teacher to an iZone school on science scores is significantly larger than the effect of other teacher turnover on science scores in the school-by-year fixed-effects models.

Previous work by Zimmer et al. (2017) found positive effects in all three iZones (Memphis, Nashville, and Chattanooga). However, the strongest and most consistent effects were in Memphis. Therefore, we conducted another set of analyses restricting the sample to only Memphis iZone’s sending schools (see Supplementary Appendix Table C1 in the online version of the journal). We find comparable results with this restricted sample.¹⁰

Heterogeneous Effects

Next, we examine the moderating effect of teachers’ prior effectiveness as measured by TVAAS scores (see Figure 1). As the results are similar between our two specifications, we display the results of this analysis using our preferred model (school-by-year fixed effect). Consistent with the previous literature (Adnot et al., 2017), the loss of effective teachers is more harmful for students. Specifically, students in grades that lost all of their reading teachers who met student growth expectations in the prior year to the iZone scored 0.16 SDs lower on their reading assessment than students in the same school and year that did not lose any effective teachers to the iZone. In math and science, these effects were even greater: 0.21 and 0.18 SDs, respectively. However, the estimate for math is statistically indistinguishable from zero. In contrast, students who lost 100% of their reading teachers who had not met expectations in the prior year actually performed slightly better in reading and math, although the results are not statistically significant. In science, the students of prior ineffective teachers who transfer also performed worse, again statistically insignificant, but less than students of effective teachers who transfer into iZone schools.

Figure 1.

Spillover effects of teacher turnover by teacher effectiveness.

Previous literature (Hanushek et al., 2016; Ronfeldt et al., 2013) has also found that teacher turnover is more harmful for students in more economically disadvantaged schools and low-performing schools. We test these hypotheses in Table 4. The first three columns display the effect estimates of losing teachers to the iZone based on the percentage of students that are economically disadvantaged in the sending school. We compare schools in the top quartile (most economically disadvantaged) and bottom quartile (least economically disadvantaged) of sending schools to the middle half of economically disadvantaged sending schools (our omitted group). In reading, students attending the most economically disadvantaged schools performed 0.17 SDs worse as a result of losing all grade-level teachers to the iZone than those in the middle half of economically disadvantaged sending schools. Our estimates are statistically indistinguishable from zero in math and science. However, this appears to be driven by imprecision of the estimates, particularly in math where the coefficients are large, but the standard errors are also large.

Table 4

Examining Moderating Effects by School Characteristics

Variable	By percent ED			By priority status
Variable	Reading	Math	Science	Reading	Math	Science
Teacher turnover to iZone	0.01	0.14	−0.13	−0.12*	−0.13	−0.17*
Teacher turnover to iZone	(0.07)	(0.21)	(0.15)	(0.05)	(0.10)	(0.07)
Teacher Turnover to iZone × Top 25% (Most Economically Disadvantaged)	−0.17*	−0.25	−0.04
	(0.08)	(0.23)	(0.16)
Teacher Turnover to iZone × Bottom 25% (Least Economically Disadvantaged)	−0.04	−0.16	0.08
	(0.08)	(0.26)	(0.15)
Teacher Turnover to iZone × Priority School				0.06	0.27	0.07
				(0.08)	(0.22)	(0.11)
R ²	0.63	0.48	0.55	0.63	0.48	0.55
N	17,419	17,048	14,414	17,419	17,048	14,414

Note. Standard errors are clustered by school and indicated in parentheses. FRPL = free or reduced-price meals; ELL = English Language Learner; ED = economically disadvantaged.

Student Controls: Gender, Race, FRPL status, Special Education status, ELL status, Mobility Status, Prior Reading Test Score, Prior Math Test Score, Prior Science Test Score.

School Controls: Percent Minority, Percent FRPL, Percent Student Mobility, School Level.

All models include school-by-year fixed effects.

†

p < .10. *p < .05. **p < .01. ***p < .001.

The remaining columns of Table 4 examine whether the effects of teacher turnover to iZones differentially affect low-performing schools. In particular, we compare sending schools that have been labeled as part of the bottom 5% of schools in the state (Priority schools) with other sending schools. Priority schools generally have larger proportions of minority and economically disadvantaged students. Whereas non–Priority sending schools are 77% minority and 75% economically disadvantaged, Priority sending schools are 98% minority and 89% economically disadvantaged. Furthermore, whereas students in non–Priority sending schools scored 0.32 to 0.50 SDs below the state average at baseline, students in Priority sending schools scored 0.84 to 1.12 SDs below average, depending on the subject. Therefore, based on previous literature, Priority sending schools should experience substantially larger negative effects than non–Priority sending schools. On the contrary, we find no differential effects for sending Priority schools relative to sending non-Priority schools. It is important to note, however, that the coefficient in math is quite large for Priority schools but is imprecisely estimated so that we cannot rule out these large effects.

Net Effects of iZone Teacher Recruitment

Given the positive effects found in previous research evaluating Tennessee’s iZone schools (Zimmer et al., 2017) and the resulting negative effects on sending schools discussed above, it is reasonable to ask what the net impact of the iZone intervention might be. However, for a number of reasons, it is difficult to directly compare the results from this analysis with results from previous work finding positive effects in iZone schools. First, the number of students in iZone schools is not the same as the number of students affected by the loss of a teacher to the iZones. Second, the types of students affected are different. As shown in Table 2, iZone schools have a greater proportion of minority and economically disadvantaged students, and many of these students have been served by one of the state’s lowest-performing schools for multiple years. Third, the effects evaluated in this study are short-term effects assessed in the year after teachers leave, while the effects evaluated in previous work spanned 1 to 3 years of the intervention. Fourth, the findings above consider the effects of all teachers who left the sending schools for iZone schools, while the positive effects of the iZone schools from prior research may only be partially attributable to the teachers hired. iZone schools also employed a number of other interventions, including changes in leadership, instructional coaching, and extending the school day, all of which may have contributed to the positive effects previously found. Finally, some of the teachers who left for iZone schools may have left their prior schools for other schools if the incentives to transfer to the iZone schools were not available.

Given these concerns, it is difficult to precisely calculate the net impact of iZone schools as a whole. Nonetheless, we account for a number of these concerns and make several assumptions to conduct an informal, back-of-the-envelope calculation of the impact of the iZone intervention net of the effects on the sending schools. In particular, we have not made a value judgment based on students’ backgrounds. In other words, the gain (or loss) in one student’s test score is not weighted any differently from the gain (or loss) of any other student’s test score. We also assume that recruiting effective teachers accounts for 80% of the positive impact in iZone schools. Previous research evaluating the iZone schools identified separate effects for each of the three cohorts of iZone schools in their first year of implementation. In each of our calculations below, we assume the smallest of the three. We also account for the number of students impacted in our estimate and assume all of the teachers who left for iZone schools would not have left in the absence of the incentives.

Figure 2 depicts the overall effect of iZone schools on reading, math, and science student achievement, taking into account the effect on both iZone schools and sending schools (on the y-axis) and the number of students impacted in each (x-axis). In Figure 2A, the positive effects at iZone schools are depicted by the striped blue blocks, and the negative effects at sending schools are depicted by the solid red blocks. Note that the negative effects displayed represent the effect after accounting for the average proportion of grade-level teachers who left.¹¹ In Figure 2B, we multiplied the size of the effect (y-axis value in Figure 2A) by the number of students affected (x-axis value in Figure 2B) and calculated the difference between these positive and negative effects to identify the net impact of the iZone schools. The figure shows a substantial positive net impact of the iZone schools in all subjects with the greatest net impact in math. Note that we used the effect estimates without regard to statistical significance.

Figure 2.

Estimation of the net effects of iZone schools.

The largest assumption we make in this comparison is that 80% of the positive impact of iZone schools is attributable to the high-quality teachers that were recruited. However, even if we assume that these teachers only explained 60% of the positive impact of iZone schools, the net impact would still be positive in all three subjects, though almost negligible in reading and science. For the negative effects on the students in sending schools to completely cancel out the positive effects on students in iZone schools, only 54% of the positive reading effect in iZone schools could be attributable to recruiting effective teachers from other schools in Tennessee. In math, this percentage would be 26%; in science, 54%.

Discussion

High-performing schools generally have a competitive advantage in the teacher labor market. These schools typically have better working conditions, less accountability pressure, and students generally viewed as easier to educate. Research shows that financial incentives have been a successful recruitment strategy for leveling the playing field and making low-performing schools more competitive in attracting high-quality teachers. States and districts across the nation are relying on effective teachers to help turnaround their lowest-performing schools. In fact, two of the three federally approved reform strategies (that allow schools to continue to operate) require or, in practice, result in the replacement of at least half of the teaching staff. However, teachers may positively affect the students in receiving schools after they transfer but the turnover may negatively affect students in the sending schools.

This study examined the spillover effects of teacher recruitment into Tennessee’s iZone schools on the students in sending schools. Although there is some variation across subjects and models, the estimates are consistently negative. Estimates range from a −0.04 to −0.12 SD change in student test score gains after taking into account average teacher turnover rates in these schools. Overall, these effects are greater than those found in prior research on teacher turnover (Hanushek et al., 2016; Ronfeldt et al., 2013), which found negative effects of 0.04 to 0.11 SDs assuming full grade-level teacher turnover. (For comparison, assuming full grade-level turnover, this study found negative effects ranging from 0.07 to 0.19 SD.) These differences may be explained by worse compositional effects in that the iZone’s recruitment of teachers focused on hiring more high-quality teachers, while prior studies examined the effect of teacher turnover regardless of teacher effectiveness. In fact, when focusing on turnover among effective teachers, the magnitude of negative effects increased in all cases, suggesting that programs only recruiting higher-performing teachers can result in even greater harm for students in sending schools.

Furthermore, if high-quality teachers were recruited from sending schools with large populations of economically disadvantaged populations or low-performing students, the unintended consequences of the teacher recruitment strategies could be more harmful than productive. Consistent with previous literature, our analysis in the case of Tennessee’s iZones finds that schools in the top quartile of economically disadvantaged students (most economically disadvantaged) suffered greater losses than the middle half, particularly in reading where students in the most economically disadvantaged schools performed 0.17 SDs worse.

Particularly relevant to the Tennessee context are Priority schools, the state’s lowest performing 5% of schools. Roughly, a quarter of sending schools were also Priority schools. If iZone schools were simply recruiting the best teachers from other Priority schools, the teacher recruitment strategy could be counter-productive if the sending Priority schools then perform even worse without these teachers. Our results, however, suggest that within Priority sending schools, grades losing teachers did not perform any better or worse than grades that did not lose teachers to the iZone. This finding could be partially explained by floor effects—the performance of Priority schools may have been so low that they cannot perform much worse. Nonetheless, it does not appear that the students in state’s lowest-performing schools are adversely affected due to the loss of their highest quality teachers to the iZone.

Our results lead us to conclude that any gains that iZone schools may be experiencing from the recruitment of high-quality teachers are being partially offset by weaker performance in the sending schools. In an informal, net-effect calculation, we compare the positive effect previously found in iZone schools to the negative effect in sending schools to estimate the net impact of the iZone initiative. Under reasonable assumptions, we find that there is a net positive effect in all three subjects. Thus, while some students experienced achievement losses, the gains acquired overall counterbalanced those losses. However, it is important to note that in this case of Tennessee’s iZone schools, only 60% of successfully recruited teachers were previously deemed effective. If future teacher recruitment programs only hire teachers who have previous records of effectiveness, the negative effects on students in sending schools could be much greater, potentially outweighing the positive effects in receiving schools and resulting in a net negative effect. In fact, while iZone schools focused on recruiting and retaining high-quality teachers in their initial years of turning around low-performing schools, the Memphis iZone eventually changed course. After several years of recruiting already-developed talent, leaders of the Memphis iZone recognized the negative impact of siphoning off teachers from non-iZone schools, depleting the number of high-quality teachers in sending schools as the iZone expanded (Glazer et al., 2020). Instead, they transitioned to a capacity building strategy focused on increasing the professional knowledge and skill of teachers and leaders already in iZone schools (Glazer et al., 2020). However, perhaps linked to this change in strategy, Pham and colleagues (2020) examined the long-term effects of the iZones and found that growth in student learning outcomes slowed in these later years, although it is not yet clear whether this transition was responsible for the diminished effects on student achievement.

Thus, a final recommendation from our study is that future research should examine long-term effects of incentives for teachers to transfer into low-performing schools. Although schools may experience a loss in achievement gains in the year immediately after a teacher exits the school, schools may be able to recover over time. For instance, sending schools may be able to hire effective replacement teachers or further develop the teachers who remain.¹² For higher-performing schools, these options may be more likely, making the recovery period rather short. If long-term analyses reveal that sending schools are able to rebound quickly, there may be even more reason to support incentives for recruiting highly effective teachers into low-performing schools, but if these policies deplete the number of effective teachers in sending schools such that no more effective teachers are willing to transfer or few effective teachers remain in sending schools, this approach may not be sustainable in the long-run, only creating an illusion of school improvement.

Supplemental Material

sj-docx-1-epa-10.3102_01623737221111807 – Supplemental material for Spillover Effects of Recruiting Teachers for School Turnaround: Evidence From Tennessee

Supplemental material, sj-docx-1-epa-10.3102_01623737221111807 for Spillover Effects of Recruiting Teachers for School Turnaround: Evidence From Tennessee by Adam Kho, Gary T. Henry, Lam D. Pham and Ron Zimmer in Educational Evaluation and Policy Analysis

Footnotes

Acknowledgements

We wish to thank the Achievement School District and Tennessee Department of Education for their feedback and assistance. Any opinions or errors are solely attributable to us, however. This study was approved by the Vanderbilt Institutional Review Board.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

Funding for this research was provided by John and Laura Arnold Foundation and the Walton Family Foundation.

ORCID iDs

Adam Kho

Lam D. Pham

Notes

Authors

ADAM KHO is an assistant professor of education policy and leadership in the Rossier School of Education at the University of Southern California. His research interests include education policies and evaluation of programs serving traditionally underserved students, with a focus on school improvement, reform, choice, and turnaround.

GARY T. HENRY is dean of the University of Delaware’s College of Education and Human Development and professor in the School of Education and the Joseph R. Biden, Jr. School of Public Policy and Administration. He specializes in education policy, educational evaluation, teacher and leader quality research, and quantitative research methods.

LAM D. PHAM is an assistant professor of educational evaluation and policy analysis in the College of Education at North Carolina State University. His research interests include school improvement and reform, with a focus on teachers and leaders in low-performing schools.

RON ZIMMER is a professor in and director of the Martin School of Public Policy and Administration at the University of Kentucky. His research interests include school choice, school finance, and school reforms. His current research focuses on school turnaround efforts and the effectiveness of school voucher programs.

References

Aaronson

Barrow

Sander

(2007). Teachers and student achievement in the Chicago public high schools. Journal of Labor Economics, 25(1), 95–135.

Adnot

Dee

Katz

Wyckoff

(2017). Teacher turnover, teacher quality, and student achievement in DCPS. Educational Evaluation and Policy Analysis, 39(1), 54–76.

Bryk

Schneider

(2002). Trust in schools: A core resource for improvement. Russell Sage Foundation.

Burnette

II . (2017). A passion for saving schools on the brink. Education Week. https://leaders.edweek.org/profile/sharon-griffin-chief-of-schools-school-turnaround/

Chetty

Friedman

J. N.

Rockoff

J. E.

(2014). Measuring the impacts of teachers II: Teacher value-added and student outcomes in adulthood. The American Economic Review, 104(9), 2633–2679.

Clotfelter

C. T.

Glennie

E. J.

Ladd

H. F.

Vigdor

J. L.

(2008). Teacher bonuses and teacher retention in low-performing schools: Evidence from the North Carolina $1,800 teacher bonus program. Public Finance Review, 36(1), 63–87.

Cowan

Goldhaber

(2015). Do bonuses affect teacher staffing and student achievement in high-poverty schools? Evidence from an Incentive for National Board Certified Teachers in Washington State. Economics of Education Review, 65, 138–152.

Dixon

L. L.

Pham

L. D.

Henry

G. T.

Corcoran

S. P.

Zimmer

(2021). Who leads turnaround schools? Characteristics of principals in Tennessee’s Achievement School District and innovation zones. Educational Administration Quarterly, 58, 258–299.

Glazer

J. L.

Egan

(2016). The Tennessee Achievement School District: Race, history, and the dilemma of public engagement. https://peabody.vanderbilt.edu/TERA/files/ASD_Dilemma_of_Engagement.pdf

10.

Glazer

J. L.

Massell

Winchell Lenhoff

Larbi-Cherif

Egan

Taylor

J. E.

Ison

Deleveaux

Millington

(2020). District-led school turnaround: Aiming for ambitious and equitable instruction in Shelby County’s iZone. https://repository.upenn.edu/cgi/viewcontent.cgi?article=1116&context=cpre_researchreports

11.

Glazerman

Max

(2011). Do low-income students have equal access to the highest-performing teachers? (No. B5869da6f9a54ef09fcd97be781b8a33). Mathematica Policy Research.

12.

Glazerman

Protik

Teh

B. R.

Bruch

Max

(2013). Transfer incentives for high-performing teachers: Final results from a multisite randomized experiment [NCEE 2014-4004]. National Center for Education Evaluation and Regional Assistance.

13.

Goldhaber

Gross

Player

(2011). Teacher career paths, teacher quality, and persistence in the classroom: Are public schools keeping their best? Journal of Policy Analysis and Management, 30(1), 57–87.

14.

Goldhaber

Lavery

Theobald

(2015). Uneven playing field? Assessing the teacher quality gap between advantaged and disadvantaged students. Educational Researcher, 44(5), 293–307.

15.

Grissom

J. A.

Loeb

Nakashima

N. A.

(2014). Strategic involuntary teacher transfers and teacher performance: Examining equity and efficiency. Journal of Policy Analysis and Management, 33(1), 112–140.

16.

Guin

(2004). Chronic teacher turnover in urban elementary schools. Education Policy Analysis Archives, 12(42), 42.

17.

Hanushek

E. A.

(2011). The economic value of higher teacher quality. Economics of Education Review, 30(3), 466–479.

18.

Hanushek

E. A.

Rivkin

S. G.

(2010). Constrained job matching: Does teacher job search harm disadvantaged urban schools? (No. w15816). National Bureau of Economic Research.

19.

Hanushek

E. A.

Rivkin

S. G.

Schiman

J. C.

(2016). Dynamic effects of teacher turnover on the quality of instruction. Economics of Education Review, 55, 132–148.

20.

Hardy

(2013). Big gains sought in student achievement at Hamilton County’s five iZone schools. https://www.timesfreepress.com/news/local/story/2013/jul/22/izone-commitment/113884/?bcsubid=baee2d42-48ae-46ff-b953-23caf84ae32b&pbdialog=reg-wall-login-created-tfp

21.

Henry

Zimmer

Attridge

Kho

Viano

(2014). Teacher and student migration in and out of Tennessee’s Achievement School District. Tennessee Education Research Alliance. https://peabody.vanderbilt.edu/research/tnedresearchalliance/files/ASD_Teacher_Student_Migration.pdf

22.

Henry

Zimmer

Kho

Pham

(2017). Recruitment and retention of teachers in Tennessee’s Achievement School District and iZone Schools. Tennessee Education Research Alliance. https://peabody.vanderbilt.edu/research/tnedresearchalliance/files/Teacher_Retention_in_ASD_and_iZone_vF.pdf

23.

Henry

G. T.

Pham

Kho

Zimmer

(2020). Peeking into the black box of school turnaround: A formal test of mediators and suppressors. Educational Evaluation and Policy Analysis, 42(2), 232–256.

24.

Henry

G. T.

Redding

(2020). The consequences of leaving school early: The effects of within-year and end-of-year teacher turnover. Education Finance and Policy, 15(2), 332–356.

25.

Isenberg

Max

Gleason

Potamites

Santillano

Hock

Hansen

(2013). Access to effective teaching for disadvantaged students [NCEE 2014-4001]. Institute of Education Sciences.

26.

Jackson

C. K.

(2018). What do test scores miss? The importance of teacher effects on non–test score outcomes. Journal of Political Economy, 126(5), 2072–2107.

27.

Jackson

C. K.

Rockoff

J. E.

Staiger

D. O.

(2014). Teacher effects and teacher-related policies. Annual Review of Economics, 6(1), 801–825.

28.

Koedel

Betts

J. R.

(2007). Re-examining the role of teacher quality in the educational production function (pp. 2007-03). National Center on Performance Incentives, Vanderbilt, Peabody College.

29.

Loeb

Kalogrides

Béteille

(2012). Effective schools: Teacher hiring, assignment, development, and retention. Education Finance and Policy, 7(3), 269–304.

30.

Murnane

R. J.

(1984). Selection and survival in the teacher labor market. The Review of Economics and Statistics, 66, 513–518.

31.

Papay

Hannon

(2018). The effects of school turnaround strategies in Massachusetts [Working Paper]. Brown University.

32.

Pham

L. D.

Henry

G. T.

Kho

Zimmer

(2020). Sustainability and maturation of school turnaround: A multiyear evaluation of Tennessee’s Achievement School District and Local Innovation Zones. AERA Open, 6(2), 1–11.

33.

Rivkin

S. G.

Hanushek

E. A.

Kain

J. F.

(2005). Teachers, schools, and academic achievement. Econometrica, 73(2), 417–458.

34.

Rockoff

J. E.

(2004). The impact of individual teachers on student achievement: Evidence from panel data. The American Economic Review, 94(2), 247–252.

35.

Ronfeldt

Loeb

Wyckoff

(2013). How teacher turnover harms student achievement. American Educational Research Journal, 50(1), 4–36.

36.

Sanders

W. L.

Wright

S. P.

Horn

S. P.

(1997). Teacher and classroom context effects on student achievement: Implications for teacher evaluation. Journal of Personnel Evaluation in Education, 11(1), 57–67.

37.

Sass

T. R.

Hannaway

Figlio

D. N.

Feng

(2012). Value added of teachers in high-poverty schools and lower poverty schools. Journal of Urban Economics, 72(2), 104–122.

38.

Springer

M. G.

Swain

W. A.

Rodriguez

L. A.

(2016). Effective teacher retention bonuses: Evidence from Tennessee. Educational Evaluation and Policy Analysis, 38(2), 199–221.

39.

Steele

J. L.

Murnane

R. J.

Willett

J. B.

(2010). Do financial incentives help low-performing schools attract and keep academically talented teachers? Evidence from California. Journal of Policy Analysis and Management, 29(3), 451–478.

40.

Steele

J. L.

Pepper

M. J.

Springer

M. G.

Lockwood

J. R.

(2015). The distribution and mobility of effective teachers: Evidence from a large, urban school district. Economics of Education Review, 48, 86–101.

41.

Sullivan

(2013). One to watch: Memphis iZone. The New Teacher Project. https://tntp.org/blog/post/one-to-watch-memphis-izone

42.

Swain

W. A.

Rodriguez

L. A.

Springer

M. G.

(2019). Selective retention bonuses for highly effective teachers in high poverty schools: Evidence from Tennessee. Economics of Education Review, 68, 148–160.

43.

Tennessee Department of Education. (2013). Tennessee Department of Education to grant signing and retention bonuses to highly effective teachers. http://www.tennessee.gov/news/34146

44.

United States Department of Education. (2009). Race to the top program executive summary. https://www2.ed.gov/programs/racetothetop/executive-summary.pdf

45.

United States Department of Education. (2010). 2010 SIG final requirements. https://www2.ed.gov/programs/sif/index.html

46.

Zimmer

Henry

Kho

(2017). The effects of school turnaround in Tennessee’s achievement school district and innovation zones. Educational Evaluation and Policy Analysis, 39(4), 670–696.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB