January 2018

How has iZone Teacher Recruitment
Affected the Performance of Other
Schools?
Examining Systemic Effects of Teacher Recruitment for School Turnaround

Adam Kho, Gary T. Henry, Ron Zimmer, Lam Pham

About this Research Brief
This research brief continues the Tennessee Education Research Alliance examination of
Tennessee’s efforts to turn around its lowest performing schools. The brief reviews the
extent to which teacher recruitment of Innovation Zones throughout the state of
Tennessee have impacted the schools from which these teachers came. It follows
previous work which looked at the impacts of Tennessee’s turnaround approaches on
student achievement and teacher recruitment.
Funding for this research was provided by the Laura and John Arnold Foundation and
the Walton Family Foundation. The content is solely the responsibility of the authors
and does not necessarily represent the official views of the foundations.

vu.edu/TNEdResearchAlliance

615.322.5538

tned.research.alliance@vanderbilt.edu

@TNEdResAlliance

 Introduction
In 2012-13, Tennessee began a series of initiatives
to turnaround its lowest-performing schools,
known as Priority schools. Among them, local
Innovation Zones (iZones) in Memphis, Nashville,
and Chattanooga have been acclaimed as some of
the most effective initiatives in improving the
academic achievement of students (Zimmer, Kho,
Henry & Viano, 2015; Gonzales, 2016; Kebede,
2016; Tillery, 2017; Zimmer, Henry & Kho, 2017).
A key strategy that contributed to the success of the
iZone schools was the recruitment of highly
effective teachers.

Key Findings
1. Students in schools losing teachers to
iZone schools experienced a small
negative effect as a result of those
teacher losses, particularly in reading
and science.
2. This effect of losing teachers was
smaller in Priority schools than in nonPriority schools.
3. The small negative effects of losing
teachers in the schools that lost these
teachers does not appear to offset the
positive effects in iZone schools.
Overall, considering both, our
calculations show a net positive impact
on students in Tennessee as a result of
the iZone initiative.

In a previous policy brief, we found that teachers
entering iZone schools averaged higher value-added
rating scores than teachers entering other lowperforming schools and other Tennessee schools in
general (Henry, Zimmer, Kho & Pham, 2017).
However, while students in the iZone schools have
likely benefitted from the instruction of these highquality teachers, it is important to examine possible
unintended consequences of this teacher
recruitment. In this policy brief, we examine the
systemic effects of the recruitment of high-quality
teachers to iZone schools. In particular, what
impact has this recruitment had on student
achievement in the schools that lost these teachers
(henceforth referred to as sending schools)? This
information will help to provide a more complete
picture of the overall effect of the iZone initiative.

demographic characteristics, and state assessment
scores. For the analysis, we use reading, math, and
science scale scores and convert these to
standardized units by subject, grade, year, and, for
EOCs, semester to have a common metric across
grades and years. Data for this analysis span school
years from 2011-12 through 2014-15.

Data

Who Was Affected?

To examine the systemic effects of the recruitment
of teachers into iZone schools, we utilized a
database provided by the Tennessee Department of
Education and compiled by the Tennessee
Education Research Alliance. The database
contains data on each teacher in Tennessee
including his/her school of employment and the
tested grades and courses he/she taught.

Between the 2012-13 and 2014-15 school years, 652
teachers transferred into one of 26 iZone schools in
Memphis, Nashville, or Chattanooga. In this
analysis, we follow the 234 teachers who previously
taught tested grades and subjects – 181 teachers
who entered iZone schools in Memphis, 28 teachers
who entered iZone schools in Nashville, and 25
teachers who entered iZone schools in Chattanooga.
Most of these teachers, about 92%, were recruited
from other schools in the same district as the
receiving iZone school they transferred to. About

The database also includes student-level data,
including school and grade enrollment data,

2

 Why Teacher Turnover May Adversely Affect Students in Sending Schools
Teacher turnover can harm students in a number of ways. Here, we highlight three mechanisms as
identified in the previous research literature (Henry & Redding, 2017) – change in the composition of
teachers, classroom disruptions, and staff instability.
Change in the Composition of Teachers. Teachers who leave a school must be replaced. If they are
replaced with a more effective teacher, students may benefit from the change in composition. However,
teacher recruitment in iZone schools specifically targeted highly effective teachers, meaning that sending
schools likely lost some of their more effective teachers. As a result, these teachers were likely to have
been replaced with less effective teachers, on average. If so, the compositional change will decrease
overall teacher quality and may decrease student achievement.
Classroom Disruptions. Teacher turnover can create classroom disruptions, particularly with midyear
transitions. Students become accustomed to the routines and procedures of their teachers and must
adjust to new routines and procedures when those teachers leave. Further, highly effective teachers are
more likely to have a set of routines and procedures for their classrooms (Emmer & Stough, 2001; Oliver
& Reschly, 2007). Therefore, losing a highly effective teacher can create greater classroom disruption.
Staff Instability. Teachers have knowledge about the schools where they work and develop relationships
with other teachers, faculty members, and students. When teachers leave, these relationships are severed
and their institutional knowledge leaves with them. The instability that results can be even greater with
the loss of highly effective teachers, as they often have more responsibilities or more often serve in
leadership roles.

4% came from nearby or bordering districts, and 3%
came from other districts throughout Tennessee.
The three districts primarily placed elementary and
middle schools into the iZones, and therefore, most
teacher transfers were from elementary and middle
schools (see Table 1). Overall, the sending schools
had smaller percentages of minority and
economically disadvantaged students than iZone
schools. Nevertheless, the sending schools, on
average, had a larger share of economically
disadvantaged students and performed worse on the
state’s standardized assessments relative to the
average Tennessee school. About a quarter of the
sending schools were also Priority schools (the
state’s lowest-performing 5% of schools).
Table 1 provides the average proficiency rates of
schools in reading, math, and science prior to iZone
initiatives. Both iZone schools and sending schools
were significantly lower-performing relative to the

state prior to the recruitment of teachers for the
iZones. Schools served by the iZone initiatives
performed worse than the sending schools in the
year prior to the implementation of the iZones.
Research Approach
For the primary analysis, we use a methodology that
allows us to compare student test score gains in
grades that lost a teacher(s) to the iZone to other
grades in the same school in the same year that did
not lose a teacher(s) to the iZone, known as a
school-by-year fixed effect approach. This
approach allows us to control for other school-level,
year-specific changes that may have affected student
achievement gains. We also account for the effect of
other teacher turnover in which the teachers did not
leave to iZone schools to avoid misattributing the
effect of turnover into the iZone schools to turnover
that occurred for other reasons.
3

 Table 1. Characteristics of iZone Schools, Sending Schools, and Tennessee Schools Prior to iZone
Interventions
Sending Schools

iZone Schools

Tennessee+

140

26

1,614

Elementary

36%

46%

56%

Middle

44%

39%

25%

High

20%

15%

19%

Percent Minority

83%

97%

33%

Percent Economically Disadvantaged

79%

91%

61%

Priority Schools (Lowest-Performing)

26%

100%

5%

Reading Average Proficiency Rates

26%

14%

49%

Math Average Proficiency Rates

30%

18%

48%

Science Average Proficiency Rates

30%

19%

60%

Characteristics
Total Schools

+

Excludes alternative schools

We present overall impact estimates of teacher
turnover to iZone schools separately for reading,
math, and science. Note that the effects displayed
are for 100% turnover of the particular grade in the
sending school.
Results
Overall Effects on Sending Schools
We find negative effects of teacher turnover to
iZone schools in all three subjects (see Figure 1). In
reading, students entering grades in which teachers
left for iZone schools experienced a decrease in
standardized test scores of about a tenth of a
standard deviation on average. In science, students
entering grades in which the teachers left for iZone
schools experienced a decrease in standardized test
scores of about 0.14 standard deviations on average.
The results in both subjects are statistically
significant. In math, the negative effect is smaller at
0.07 and not statistically significant. To give context
and help interpret this metric, the average difference
between the cutoffs for basic and proficient

achievement levels on the 2013-14 state assessment
was 1.28 standardized units in reading, 1.19 in
math, and 1.24 in science. Given this context, the
results can be generally viewed as relatively small
adverse effects for sending schools, especially
considering these estimates represent the effect of
all teachers leaving for the iZone schools in the
grade and school where the turnover occurred. On
average, in the grades of sending schools in which
turnover occurred, 46% of the grade-level reading
teachers left for iZone schools. In math and science,
the iZone-bound turnover averaged 54% and 58%,
respectively. Therefore, the average effect on
students in grades within schools that sent a teacher
to the iZone was about half of the effect described
above or approximately 0.05 standard deviation
units for reading or 0.08 standard deviation units
for science, spread across all the students within the
grade.
Effects on Sending Schools by Priority School Status
On average, sending schools were higherperforming than iZone schools. However, about a

4

 Figure 1. Estimates of the Effects of Teacher
Turnover on Student Achievement in the Schools
Sending Teachers to iZone Schools
0.30
0.20
0.10
0.00
-0.10

-0.10*

-0.07

Reading

Math

-0.14*

-0.20
-0.30
Science

Note: An asterisk (*) indicates statistical significance of estimates at
the 0.05 level

quarter of sending schools were also Priority
schools (see Table 1), meaning the teachers who left
these schools left a low-performing environment to
transfer to another low-performing en vironment.
Prior research shows that teacher turnover is more
harmful for lower-achieving schools (Ronfeldt, Loeb
& Wyckoff, 2013; Hanushek, Rivkin & Schiman,
2016). Therefore, for these schools, the unintended
consequence of the iZone teacher recruitment could
be even more harmful to the students in the sending
schools.
To assess this concern, we disaggregated our
analysis on sending schools into Priority schools
and non-Priority schools. We found that the effects
in Priority schools appear to be less than other
schools and are not statistically significant.
Assessing Overall iZone Effects
Given the positive effects found in iZone schools
and the resulting negative effect in sending schools,
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 the results from
our previous policy brief highlighting the effect of
iZone schools. First, the number of students

impacted by the iZone school initiatives is not the
same as the number of students impacted by the
loss of a teacher to an iZone school. Second, the
types of students affected are different. As shown in
Table 1, 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 policy
brief are short-term effects assessed in the year after
teachers leave. The effects evaluated in the previous
brief spanned one to three years of the intervention.
Lastly, this analysis only considers the effects of
teachers at the sending schools. The positive effects
of the iZone 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 we found.
Given these concerns, it is difficult to precisely
calculate the net impact of iZone schools as a whole.
Nonetheless, we take into account a number of
these concerns and make several assumptions here
to conduct a less formal, back-of-the-envelope
calculation of the net impact of iZone 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. In our previous
policy brief, we identified separate effects for each of
three cohorts of iZone schools in their first year. 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 below.
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

5

 Figure 2. Estimation of the Net Effects of iZone Schools
(a)
0.25

Reading

Math

Science

0.20

Effect Size

0.15
0.10
0.05
0

1 2 3 4 5 6 7 8

1 2 3 4 5 6 7 8

1 2 3 4 5 6 7 8

Students
(1000’s)

-0.05
-0.10

Positive Effect on iZone Schools
Negative Effect on Sending
Schools
(b)

Effect Size x Students

1000
800
600
400
200

Reading

Math

Science

Net Effect on Tennessee Schools

6

 and sending schools1 and the number of students
impacted in each. In Figure 2a, the positive effects
at iZone schools are displayed in blue, and the
negative effects at sending schools are displayed in
red. In Figure 2b, we multiplied the size of the
effect by the number of students affected and found
the difference between these positive and negative
effects to identify the net impact of the iZone
schools, displayed in green. This figure shows a
large 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.
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 half 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 students in sending schools to completely
cancel out the positive effects on students in iZone
schools, only 49% of the positive science effect in
iZone schools could be attributable to recruiting
effective teachers from other schools in Tennessee.
In reading, this percentage would be 47%; in math,
23%.
Conclusion
In 2012-13, Tennessee began a series of
interventions to improve the academic performance
of low-performing schools throughout the state.
Prior research shows that district-led iZones were
effective in increasing students’ achievement gains.
A key strategy that contributed to the success of the
iZones was the recruitment of highly effective
teachers – the iZone schools hired 652 teachers

1

Note that the effects displayed in Figure 1 represent the effect
as a function of full grade-level turnover. The effect sizes
depicted in Figure 2 represent the full effect displayed in

from throughout the state to help turnaround these
schools.
In this research brief, we examine possible negative
side effects of recruiting highly effective teachers
into the iZone schools – did students in sending
schools have to lose in order for students in iZone
schools to gain? Our analysis shows that the
students in the sending schools did experience a
small loss in terms of gains on achievement tests. In
grades that lost a teacher to an iZone school,
students averaged losses in test score gains between
0.04 to 0.08 standard deviations, depending on the
subject, due to this teacher transition.
However, this decrease in student performance at
sending schools appears to be small relative to the
positive effect of iZone schools. A direct
comparison and therefore net effect estimation
cannot be made using our previous findings of the
impact of iZones on student achievement and the
impact of teacher turnover on sending schools to
iZone schools in this analysis. We can, however,
make an informal calculation of an overall net effect
of these two analyses. This informal calculation
indicates a positive net impact of the iZone schools
on students’ overall achievement gains.
Our findings suggest that the iZones’ strategy for
recruiting highly effective teachers to turnaround
their schools has been successful in improving the
achievement of students in persistently lowperforming schools with a relatively small negative
side effect on the sending schools, at least in the
short run. However, if sending schools recover
from this initial dip in the longer run, this small
negative consequence may be minimized, while the
iZone schools posted consistent gains two and three
years after the intervention began. This is a
question we hope to investigate in a future report.

Figure 1 multiplied by the average proportion of grade-level
teachers that left. Therefore, the effects in Figure 2 represent
the negative effect for the average grade that lost teachers.

7

 References
Emmer, E. T., & Stough, L. M. (2001). Classroom management: A critical part of educational psychology, with
implications for teacher education. Educational psychologist, 36(2), 103-112.
Gonzales, J. (2016, February 17). Memphis iZone schools outpaced state-run turnaround efforts. The
Tennessean. Retrieved from http://www.tennessean.com/story/news/education/2016/02/17/memphis-izoneschools-outpace-state-run-turnaround-efforts/78693182/.
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.
Henry, G. & Redding, C. (Working Paper). The Consequences of Leaving School Early: The Effects of WithinYear and End-of-Year Teacher Turnover.
Henry, G., Zimmer, R., Kho, A. & Pham, L. (2017). Recruitment and Retention of Teachers in Tennessee’s
Achievement School District and iZone Schools. Retrieved from https://peabody.vanderbilt.edu/research/
tnedresearchalliance/files/Teacher_Retention_in_ASD_and_iZone_vF.pdf.
Kebede, L. F. (2016, May 2). iZone chief Sharon Griffin on fixing Memphis’ most challenging schools.
Chalkbeat. Retrieved from http://www.chalkbeat.org/posts/tn/2016/05/02/izone-chief-sharon-griffin-on-fixingmemphis-most-challenging-schools/.
Oliver, R. M., & Reschly, D. J. (2007). Effective Classroom Management: Teacher Preparation and Professional
Development. TQ Connection Issue Paper. National Comprehensive Center for Teacher Quality.
Ronfeldt, M., Loeb, S., & Wyckoff, J. (2013). How teacher turnover harms student achievement. American
Educational Research Journal, 50(1), 4-36.
Tillery, A. (2017, January 17). The iZone, How School Districts are Closing the Achievement Gap. Ebony.
Retrieved from http://www.ebony.com/news-views/izone-schools-memphis-philadelphia#axzz4dEDmiuzp.
Zimmer, R., Henry, G. & Kho, A. (2017). The Effects of School Turnaround in Tennessee’s Achievement School
District and Innovation Zones. Educational Evaluation and Policy Analysis, 39(4), 670-696.
Zimmer, R., Henry, G., Kho, A. & Viano, S. (2015). Evaluation of the Effect of Tennessee’s Achievement School
District on Student Test Scores. Retrieved from https://peabody.vanderbilt.edu/research/tnedresearchalliance/
files/ASD_Impact_Policy_Brief_Final_12.8.15.pdf.

8

vu.edu/TNEdResearchAlliance

615.322.5538

tned.research.alliance@vanderbilt.edu

@TNEdResAlliance