Teaching and Teacher Education 67 (2017) 330e339

Contents lists available at ScienceDirect

Teaching and Teacher Education

journal homepage: www.elsevier.com/locate/tate

How principal leadership and person-job fit are associated with

teacher mobility and attrition
Daniel Player a, *, Peter Youngs b, Frank Perrone c, Erin Grogan d
a
Batten School of Leadership and Public Policy, University of Virginia, 235 McCormick Road, Charlottesville, VA 22904, United States
b
Department of Curriculum, Instruction, & Special Education, University of Virginia, 405 Emmet Street South, Charlottesville, VA 22904, United States
c
Department of Teacher Education, Education Leadership & Policy, University of New Mexico, Albuquerque, NM 87131, United States
d
TNTP, 186 Joralemon St., Suite 300, Brooklyn, NY 11201, United States

h i g h l i g h t s

Leadership fit is associated with 35 percent lower odds of leaving the profession.

Person-job fit is associated with 65 percent lower odds of leaving teaching.

No observable characteristics moderate principal leadership and mobility.

a r t i c l e i n f o a b s t r a c t

Article history: While existing studies of teacher retention have attempted to isolate economic and organizational fac-
Received 6 July 2016 tors that predict teacher turnover, this paper extends the research base by incorporating measures of
Received in revised form principal leadership and person-job (P-J) fit. Using data from roughly 3000 teachers from the 2011-12
30 May 2017
Schools and Staffing Survey and the 2012-13 Teacher Follow-up Survey, we explore how leadership and
Accepted 17 June 2017
Available online 15 July 2017
P-J fit are associated with teachers' mobility. The results confirm that leadership and P-J fit predict
retention in one's school and retention in the teaching profession, respectively, and we find no evidence
that these associations are moderated by school or teacher characteristics.
Keywords:
Teacher retention
© 2017 Elsevier Ltd. All rights reserved.
Fit
Teacher organization fit
Teacher school fit
Teacher attrition
School leadership

The ability of a school to retain a sufficient number of high- 2011).

performing teachers has a significant impact on its functioning
over time (e.g., Ingersoll & May 2012; Ronfeldt, Loeb, & Wyckoff,
2013). Research on teacher retention has typically focused on
how student demographics and teacher characteristics have
affected teachers’ decisions to remain in their schools, move to
other schools, or leave teaching (e.g., Clotfelter, Ladd, Vigdor, &
Diaz, 2004; Hanushek, Kain, & Rivkin, 2004; for reviews of this
literature, see Borman & Dowling, 2008; Guarino, Santibanez, &
Daley, 2006). Several studies have also examined how principal
leadership affects teacher retention (e.g., Allensworth, Ponisciak, &
Mazzeo, 2009; Boyd et al., 2011a; Ingersoll & May 2012; Ladd,
* Corresponding author.
E-mail address: dplayer@virginia.edu (D. Player).
But perhaps surprisingly, there has been relatively little focus in
research on K-12 education on the role of person-environment fit in
explaining teacher turnover. Theories of fit emerged from a robust
research base in industrial organizational (I-O) psychology that has
explored how employees interact with their work environments in
an attempt to understand factors that lead to retention and other
desirable outcomes (Chatman, 1989; Kristof-Brown, Zimmerman, &
Johnson, 2005). Meta-analytic reviews of I-O research have found
moderate associations between employee retention and different
types of person-environment fit (Hoffman & Woehr, 2006; Kristof-
Brown et al., 2005), although this is one of the first studies to
examine actual turnover rather instead of stated intent to leave.
While there are several ways to conceptualize fit with regard to
work environments, this paper uses two nationally representative
datasets from the U.S., the 2011-12 Schools and Staffing Survey

http://dx.doi.org/10.1016/j.tate.2017.06.017
0742-051X/© 2017 Elsevier Ltd. All rights reserved.
 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339
(SASS) and the 2012-13 Teacher Follow-up Survey (TFS), to inves-
tigate how fit with the demands of the job (i.e., person-job fit) and
principal leadership relate to teacher retention. In our analyses, we
found that principal leadership strongly predicted retention in
one's school while person-job (P-J) fit strongly predicted retention
in the teaching profession and retention in one's school. These
findings are important because they indicate that school districts
should attend to leadership practices in schools and teaching ap-
plicants' match with the demands of the profession in order to
increase the likelihood that teachers are retained over time.
In the first two sections of this paper, we review previous
research on principal leadership, person-job fit, and teacher
retention. In the third section, we present a series of hypotheses
regarding leadership, P-J fit, and retention that we tested in our
analyses. The fourth section describes our methods including our
data and sample, measures, and model. In the fifth section, we
present our main findings regarding the role of principal leadership
and P-J fit in teacher retention. Finally, we discuss implications of
our findings for efforts to retain teachers in their schools and in the
profession.
1. Principal leadership and teacher retention
Prior research has examined how teacher retention is affected
by student demographics, teacher characteristics, and teachers'
working conditions. Studies have reported that teachers are more
likely to leave schools that serve high percentages of low-income,
non-White, and/or low-achieving students (Lankford, Loeb, &
Wyckoff, 2002; Scafidi, Sjoquist, & Stinebrickner, 2007). In addi-
tion, teachers' age, years of experience, and effectiveness have been
found to predict turnover. For example, early career teachers and
those close to retirement are more likely to leave their positions
than mid-career teachers (Allensworth et al., 2009; Guarino et al.,
2006; Ingersoll, 2001). At the same time, teachers who are more
effective (as measured by students’ performance on state tests) are
more likely to stay in their positions than less effective teachers
(Goldhaber, Gross, & Player, 2011; Boyd et al., 2011a).
Several studies have found that strong principal leadership can
promote teacher retention even in contexts where student and
teacher characteristics predict that turnover is likely. For example,
Boyd et al. (2011a) examined the impact of six aspects of school
context on the retention of all first-year K-12 teachers in New York
City: school leadership, teachers' classroom autonomy and influ-
ence on school policies, teachers’ relations with colleagues, student
behavior, school safety, and school facilities. After controlling for
other school and teacher characteristics, the authors found that
school leadership was the only factor that significantly predicted
teacher retention. Their measure of school leadership was based on
several survey items including whether an effective school disci-
pline policy was in place, whether the school administration was
viewed as supportive, whether the school administration was
perceived as evaluating teachers fairly, and whether the school
administration consulted with faculty before making decisions. In
another study, Allensworth et al. (2009) investigated the effect of
principal leadership and other school organizational conditions on
the retention of beginning K-12 teachers in Chicago. The study re-
ported that retention was higher in schools where teachers viewed
the principal as a strong instructional leader, schools where
teachers expressed high levels of trust in their principal, and
schools where teachers reported having notable influence over
school decisions.
Ladd (2011) drew on a 2006 statewide survey administered to
all K-12 teachers in North Carolina to explore the relationship be-
tween teachers' perceptions of working conditions and their
planned and actual departure rates from their schools. These
331
working conditions included the quality of school leadership,
expanded roles for teachers, PD for teachers, facilities and re-
sources, and, at the elementary and middle school levels, time to
meet with students and other teachers. Ladd's measure of the
quality of school leadership was based on a range of survey items
including whether the principal was viewed as supportive with
regard to instruction and student discipline, whether the principal
maintained high expectations for student learning and teachers'
instruction, whether teachers trusted the principal, whether the
principal involved teachers in decision making, and, at the
elementary and middle school levels, whether teachers viewed the
teacher evaluation process as legitimate and fair. Ladd (2011) re-
ported that the quality of school leadership was a stronger pre-
dictor of teachers' planned and actual departures than any of the
other working conditions variables. In addition, the quality of
principal leadership had a stronger effect on teacher attrition than
the percentage of students eligible for free or reduced price lunch
or the percentage of racial/ethnic minority students.
Research from outside the U.S. also indicates that the nature and
quality of school leadership are strongly associated with teachers'
plans to remain at their schools. Hulpia, Devos, and Van Keer (2011)
drew on data from more than 1500 secondary teachers in 46
schools in Belgium to consider factors that affect teachers'
commitment to their schools; research indicates that commitment
is highly correlated with teacher retention (Chan, Lau, Nie Lim, &
Hogan, 2008; Ladd, 2011). The researchers found that teacher
commitment was related to the quality of support provided by
school leaders, the degree of cooperation among school leaders,
and the extent to which teachers contribute to school decision
making (Hulpia et al., 2011). In this study, the measure of sup-
portive school leadership included items that measured school
administrators’ strength of vision, supportive behavior, provision of
instructional support, and provision of intellectual stimulation.
Skaalvik and Skaalvik (2011) used data from over 2500
elementary and middle school teachers in Norway to examine how
supervisory support and other aspects of the school context
affected teacher satisfaction and commitment to the teaching
profession. Supervisory support was assessed using a three-item
scale that measured teachers' experiences receiving cognitive and
emotional support from school leaders while job satisfaction was
assessed using a four-item scale that measured teachers' general
enjoyment of and affective response to their work. The study re-
ported that supervisory support was directly related to teachers'
feeling of belonging at their schools and indirectly related to
teacher job satisfaction; the indirect relationship between super-
visory support and satisfaction was mediated by teachers’ sense of
belonging (Skaalvik & Skaalvik, 2011).
In sum, there is strong evidence across different national con-
texts that principal leadership is associated with teacher retention
and related outcomes. As described below, in this study we oper-
ationalized effective principal leadership as communicating a
vision for the school, providing support to teachers, recognizing
strong teacher performance, and enforcing rules for student
behavior.
2. Person-job fit and teacher retention
Theories of person-job fit have arisen from research in
industrial-organizational psychology exploring how people
interact with their environments (Chatman, 1989; Kristof-Brown
et al., 2005). Person-environment research, which has also been
influenced by scholarship on vocational choice and personality,
eventually expanded to consider not just interaction between
people and environments, but, explicitly, how well people fit or
match with the supplies and demands of their environment
332 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339
(Edwards & Shipp, 2007; Kristof-Brown et al., 2005; Ryan & Schmit,
1996).
Scholars of person-environment fit have generally explored
these interactions from two distinct, but parallel, conceptions of fit:
complementary fit and supplementary fit (Cable & Judge, 1996;
Edwards & Shipp, 2007). Complementary fit refers to situations
in which individuals’ characteristics (e.g., skills, abilities, or values)
fill a gap in the needs of a job, organization, or group or when the
job, organization, or group is able to meet the needs of individuals
by providing rewards or experiences which they desire (Cable &
Judge, 1996; Kristof-Brown et al., 2005). The former situation may
be thought of as “demands-abilities” fit, while the latter may be
called “needs-supplies” fit (Kristof-Brown et al., 2005; Piasentin &
Chapman, 2006). In contrast, supplementary fit describes situa-
tions in which people and organizations are similar in their char-
acteristics or, simply put, places where people view themselves as
“fitting in” (Muchinsky & Monahan, 1987, p. 270).
In this paper, we explore the role of person-job (P-J) fit and
principal leadership in predicting teacher retention. P-J fit may be
conceptualized as either the correspondence between employee
needs and job ‘supplies,’ or alternatively as a match between needs
and job ‘demands’ (Edwards, 1991). Person-job fit is generally
thought of as a manifestation of complementary fit, and, more
specifically, of demands abilities fit (Edwards & Shipp, 2007). Given
the nature of the data used in this study (described in more detail
below), we operationalize P-J fit as congruence between employee
needs and the demands of the teaching profession, although his-
torically there has been less research into the job demands
conceptualization of P-J fit (Edwards, 1991). When individuals have
the abilities required to complete the tasks of a given position,
person-job fit is said to be high (Edwards, 1991; Kristof-Brown,
Jansen, & Colbert, 2002).
Existing I-O research suggests that P-J fit is positively related to
individual performance and adjustment at work and significantly
predicts attitudes toward the organization (Caldwell & O'Reilly,
1990). Research has shown that high levels of P-J fit are positively
associated with organizational commitment and negatively asso-
ciated with intent to quit, which is particularly relevant for studies
with retention as the primary outcome (Kristof-Brown et al., 2005;
Lauver & Kristof-Brown, 2001). In addition, a growing number of
studies in K-12 education from within and outside the U.S. have
examined associations between principal leadership and P-J fit or
similar constructs, such as job satisfaction.
De Nobile and McCormick (2008) employed data from more
than 300 teachers in Catholic primary schools in New South Wales,
Australia and explored how different forms of within-school
communication affected teacher job satisfaction. They found that
downward supportive communication (i.e., supportive communi-
cation from school leaders and other more senior staff) had a
stronger association with teacher satisfaction with (a) supervision
and (b) the work of teaching than other forms of communication. In
addition, they reported that openness of communication was more
strongly associated with satisfaction with one's relationship with
their principal than other factors (De Nobile & McCormick, 2008).
Van Maele and Van Houtte (2012) used data from over 2000
secondary teachers in 80 schools in Belgium to investigate the ef-
fect of teachers’ trust relationships with their principals and
teacher colleagues on their satisfaction with their work. The au-
thors found that teacher-principal trust and teacher-teacher trust
had a significant influence on teacher satisfaction. Tickle, Chang,
and Kim (2011) drew on data from the 2003-04 SASS to examine
factors associated with teacher job satisfaction and commitment.
They found that administrative support was strongly associated
with job satisfaction and intention to remain in teaching; and that
job satisfaction also predicted plans to continue in teaching (Tickle
et al., 2011).
Few studies have focused on P-J fit and teachers' actual retention
decisions. One exception is a study by Kelly and Northrop (2015)
that used data from the Beginning Teacher Longitudinal Survey
(BTLS). In their study, the authors drew on three years of data on
1770 novice K-12 teachers in the U.S. and examined how teachers'
perceptions about school support and the teaching profession
affected their retention decisions. The authors measured school
support through four survey items on principal leadership and
three items on support from other teacher colleagues. The items on
career satisfaction (i.e., fit with the profession) included questions
about the teacher's degree of enthusiasm for teaching, whether the
challenges involved in teaching were manageable, and whether
they were likely to leave teaching for work outside of education.
The authors reported that career satisfaction had a strong influence
on retention in the teaching profession as opposed to retention in
one's school (Kelly & Northrop, 2015).
3. Hypotheses
In this paper, we examined how teacher mobility was affected
by principal leadership and P-J fit. As discussed above, several
studies have found that effective principal leadership is associated
with teachers’ plans to remain at their schools or remain in
teaching as well as their actual retention decisions (Allensworth
et al., 2009; Boyd et al., 2011a; Ladd, 2011). Drawing on this
research, we hypothesized that principal leadership would be
related to retention in the following ways:
Hypothesis 1. The greater the degree of effective principal leader-
ship, the lower the likelihood the teacher will switch schools or leave
teaching. Effective principal leadership is expected to keep a teacher in
the profession at the school where they experienced such leadership.
As noted above, research in I-O psychology has shown that high
levels of P-J fit are negatively associated with plans to leave one’s
position (Kristof-Brown et al., 2005; Lauver & Kristof-Brown, 2001).
In K-12 education, Tickle et al. (2011) reported that job satisfaction
was linked to plans to remain in teaching and Kelly and Northrop
(2015) found that career satisfaction was associated with reten-
tion in teaching. Thus, we hypothesized two ways in which P-J fit
would be related to teacher retention:
Hypothesis 2A. The lower the level of one’s P-J fit, the more a teacher
is misaligned with the profession and the more likely they are to leave
the teaching profession entirely.
Hypothesis 2B. Low P-J fit (misalignment with teaching) will not be
significantly related to switching schools; a teacher who is not well-
suited for the demands of teaching will be more likely to leave the
profession completely than to seek out a new school.
We hypothesized that school characteristics may influence
principal leadership and may explain some of the relationship be-
tween leadership and retention. In the U.S., elementary schools are
generally smaller than middle schools, in terms of the number of
students and teachers, and elementary school teachers are likely to
have more frequent direct contact with their principals than middle
school teachers. Middle schools often have one or more assistant
principals and middle school teachers may have more direct con-
tact with an assistant principal than their principal. For these rea-
sons, we hypothesized that effective principal leadership will have
a stronger effect on teacher retention at the elementary level than
the middle school level.
Hypothesis 3A. The relationship between leadership and retention
will be greater for elementary than secondary teachers.
 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339 333

Studies have found that teacher turnover is higher in charter See Table 1 for a summary of these hypotheses and the relevant
schools than it is in traditional public schools (Goldring, Taie, & studies that provide support for them.
Riddles, 2014; Stuit & Smith, 2012). For example, drawing on data
from the 2008-09 TFS, Keigher (2010) reported that charter school
4. Methods
teachers were less likely to stay in their schools of origin than
traditional public school teachers (76.2% vs. 84.6%), and they were
4.1. Data and sample
more likely to move to different schools (11.4% vs. 7.5%) and leave
teaching (12.5% vs. 7.9%). Therefore, we hypothesized that effective
Data for this study came from the 2011-12 Schools and Staffing
principal leadership would have a stronger effect on teacher
Survey (SASS) and the 2012-13 Teacher Follow-Up Survey (TFS). The
retention in charter schools than those in traditional public schools.
SASS is a very comprehensive data source for research on the or-
Hypothesis 3B. The relationship between leadership and retention ganization and staffing of elementary and secondary schools. It
will be greater in charter schools than in traditional public schools. consists of a series of linked surveys administered to district ad-
ministrators, school principals, and teachers. In this study, we used
Research indicates that teacher retention rates are similar across
data from the Public School Teacher, Principal, School, and District
urban, suburban, and rural schools (Smith & Ingersoll, 2004). Thus,
Questionnaires; we did not include data from questionnaires
we hypothesize that effective principal leadership will have com-
administered to private schools because the factors that affect
parable effects on retention across different school locations.
teacher retention in private schools are likely to differ from those
Hypothesis 3C. The relationship between leadership and retention that affect retention in public schools.
will be similar in urban, suburban, and rural schools. Data for the 2011-12 SASS and the 2012-13 TFS were collected
using a stratified probability sample design with schools sampled
Studies have found that teacher turnover is higher in schools
first, followed by school districts. Schools were selected with a
that serve high percentages of low-income students than in other
probability proportionate to the square root of the number of
schools (Boyd, Lankford, Loeb, Ronfeldt, & Wyckoff, 2011b;
teachers and they were selected to be representative at the national
Hanushek et al., 2004; Ingersoll, 2001). Therefore, we hypothe-
and state levels. The weighted school response rate was 72.5
sized that effective principal leadership would have a stronger ef-
percent. To obtain the teacher sample, principals were contacted
fect on teacher retention in high poverty schools than in other
and asked to submit a list of all teachers currently working in their
schools.
building, with a weighted response rate of 79.6 percent. From these
Hypothesis 3D. The relationship between leadership and retention lists, teachers were assigned to strata based on race/ethnicity,
will be greater in high poverty schools than in low poverty schools. assignment in a classroom where students had Limited English
Proficiency, and “beginning teacher” status (i.e., they had been
Research indicates that beginning teacher commitment tends to
teaching for 3 years or less). For each sampled school, one to 20
vary considerably within individual teachers during their first years
teachers were sampled and the weighted teacher response was 77.7
of teaching before stabilizing as some individuals leave the pro-
percent.
fession while others make longer-term plans to remain in it
The 2011-12 SASS is linked to the 2012-13 TFS, which was
(Pogodzinski, Youngs, & Frank, 2013; Coldwell, 2016). In addition,
administered 12 months after the 2011-12 Teacher Questionnaire
early career teachers tend to leave teaching at higher rates than
to a sample of teachers who had completed the first questionnaire;
more experienced teachers. Thus, we hypothesized that effective
the weighted response was 81.3 percent. The TFS was designed to
principal leadership would have a stronger effect on retention for
support comparative analysis of teachers who continued teaching
teachers in their first three years of teaching than other teachers.
in their original schools (“stayers”), teachers who remained in
Hypothesis 3E. The relationship between leadership and retention teaching, but switched schools (“movers”), and teachers who left
will be greater for novice teachers (i.e., teachers in their first 3 years in the teaching profession (“leavers”). The TFS was stratified by school
the profession) than for more experienced teachers. levels (elementary vs. secondary), experience (beginning teacher
vs. experienced), and school sector (public vs. private).

Table 1
Hypotheses tested in this study.

Hypotheses Relevant Studies Supporting Hypotheses

1. The greater the degree of effective principal leadership, the lower the likelihood the teacher will Allensworth et al. (2009); Boyd et al. (2011a) and Boyd et al.
switch schools or leave teaching. Effective principal leadership is expected to keep a teacher in the (2011b); Ladd (2011)
profession at the school where they experienced such leadership.
2A. The lower the level of one's P-J fit, the more a teacher is misaligned with the profession and the Kelly and Northrop (2015); Kristof-Brown et al. (2005); Lauver
more likely they are to leave the teaching profession entirely. and Kristof-Brown (2011); Tickle et al. (2011)
2B. Low P-J fit (misalignment with teaching) will not be significantly related to switching schools; a
teacher who is not well-suited for the demands of teaching will be more likely to leave the profession
completely than to seek out a new school.
3A. The relationship between leadership and retention will be greater for elementary than secondary
teachers.
3B. The relationship between leadership and retention will be greater in charter schools than in Goldring et al. (2014); Keigher (2010); Stuit and Smith (2012).
traditional public schools.
3C. The relationship between leadership and retention will be similar in urban, suburban, and rural Smith and Ingersoll (2004)
schools.
3D. The relationship between leadership and retention will be greater in high poverty schools than in Boyd et al. (2011b); Hanushek et al. (2004); Ingersoll (2001)
low poverty schools.
3E. The relationship between leadership and retention will be greater for novice teachers (i.e., teachers Pogodzinski et al., 2013; Coldwell (2016)
in their first 3 years in the profession) than for more experienced teachers.
334 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339
To create the final sample used in this analysis, data from the TFS
were merged with data from the SASS Public School Teacher
Questionnaire; this enabled us to determine teachers’ employment
status in 2013. Thus, the final dataset was limited to those teachers
whose 2013 employment status was known. We also restricted the
dataset to include only full-time teachers in regular public and
charter schools. Finally, we merged data from the Principal and
School Questionnaires with this dataset. The resulting dataset used
for our analysis included the full-time regular public and charter
teachers that could be matched to principal and school survey
responses.
4.2. Ethics and ethical approval
The 2011-12 SASS and the 2012-13 TFS are restricted-use ver-
sions of data provided by the National Center for Education Sta-
tistics. In working with the SASS and TFS data, the authors took
steps to store the data on password-protected computers in locked
offices as required by the data license agreement.
4.3. Measures
4.3.1. Mobility measure
In this analysis, we distinguish between complete attrition from
the teaching profession and switching schools. Thus, our depen-
dent variable is a three-category variable representing the teacher's
employment status as reported in the TFS: switching schools
(“movers”), leaving teaching (“leavers”), or remaining in the same
school (“stayers”). Other studies of teacher retention have used
similar outcome measures (Allensworth et al., 2009; Boyd et al.,
2011b; Ingersoll, 2001).
4.3.2. Principal leadership and person-job fit measures
The creation of the principal leadership measure and the P-J fit
measure relied on confirmatory factor analysis (CFA) conducted
using maximum likelihood on SASS question items hypothesized to
indicate leadership and P-J fit. Due to the Likert-scale structure of
the responses, we utilized generalized structural equation
modeling with the hypothesized survey items after directionally
recoding responses so that the positive answers corresponded with
higher ordinal values. All of the items in our final factors load above
Comrey and Lee’s (1992) threshold of 0.45 for “fair” while 7 items
exceeded Comrey and Lee's 0.55 criteria for “good” and 6 items fit
the 0.71 loading criteria for “excellent.” Our final, best-fitting model
included 9 items, 4 of which loaded onto the principal leadership
measure and 5 of which loaded onto the P-J fit measure. The only
changes to the base model were several sequential allowances for
covariance between paired items that had large modification
indices (See Fig. 1 for final CFA structure.).
The items loading onto principal leadership asked teachers how
supportive and encouraging the administration was
(loading ¼ 0.79), whether staff members are properly recognized
for good work (loading ¼ 0.81), if the principal enforces the rules
and gives disciplinary support to teachers (loading ¼ 0.76), and
whether the principal knows the type of school he or she wants
along with how well he or she communicates this to the staff
(loading ¼ 0.69). The P-J fit items questioned whether the teacher
feels: the stress and disappointments of working at the school are
worth it (loading ¼ 0.73), less enthusiastic now than at the
beginning of his or her career (loading ¼ 0.76), that he or she would
become a teacher again (loading ¼ 0.58), about wanting to stay at
home instead of teaching because he or she feels too tired
(loading ¼ 0.50), and if he or she would leave his or her job for a
higher paying one as soon as possible (loading ¼ 0.60) (See Table 2
for final loadings and goodness of fit statistics.). In keeping with Hu
and Bentler’s (1999) 2-index presentation strategy and acceptable
model fit criteria, we triangulate our goodness-of-fit statistics and
report good fit from the two factor model (c2 (20,
N ¼ 3870) ¼ 214.544, p < 0.001; RMSEA ¼ 0.050; SRMR ¼ 0.030;
CFI ¼ 0.985; and TLI ¼ 0.973).1
After establishing the leadership and P-J fit factors, factor scores
were predicted using a least squares regression approach on all of
the full-time K-12 teachers who participated in both the SASS and
TFS. Each factor was then standardized to have a mean of 0 and a
standard deviation of 1 within the sample. The two factors had a
weighted correlation of r ¼ 0.59.
4.4. Model
We used a multinomial logit model to simultaneously model the
three possible outcomes of teachers in the TFS: stay in the same
school, move to a new school, or leave teaching. The multinomial
logit is an extension of the binary logistic regression model that
allows for flexibility in modeling the probability of multiple out-
comes, where each coefficient from the model is the marginal
probability of the outcome occurring relative to the base case. Thus,
a single multinomial logit model will produce k-1 sets of co-
efficients, where k is the number of outcomes. For this case, the
multinomial logit model is modeling three outcomes (staying,
moving, or leaving) and thus produces two sets of coefficients to
represent the marginal probability of moving and leaving relative
to staying.
The multinomial logit model is specified as
Pðy ¼ jjx; t; sÞ
,"

 X
2
¼ exp x0bj þ t0gj þ s0dj 1þ expðx0bh þ t0gh þ s0dh Þ
h¼1
#
 j
¼ 1; 2
where x is a vector of the leadership and P-J fit measures (described
above), t is a vector of teacher characteristics, s is a vector of school
characteristics, and j equals zero (stay in the same school), one
(move schools), or two (leave teaching). As controls, we included
various measures to account for teacher characteristics and school
characteristics that may also influence teacher mobility. Teacher
controls included indicators for whether the teacher was in the first
three years of teaching; the number of years taught at the current
school; gender; race/ethnicity; marital status; union membership;
age; master's degree; possessing traditional (i.e., not alternative)
certification; being the same race as the majority of students, col-
leagues, and the principal, respectively; total earnings from
teaching; whether the teacher taught special education, mathe-
matics, or science; and the number of Individualized Education
Plan (IEP) and Limited English Proficient (LEP) students taught.
School characteristics included size, schoolwide Title I status,
school urbanicity, and the percentage of racial/ethnic minority
teachers and students. The simultaneous inclusion of these controls
results in the estimation of partial effects of P-J fit and leadership
holding these observable characteristics constant.
The coefficients from multinomial logit models are difficult to
interpret in their raw form. For ease of interpretation, we report the
1
Reported fit statistics were obtained from an unweighted structural equation
model.
 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339 335

Fig. 1. Final Weighted Generalized Structural Equation Model.

All factor loadings are standardized.

Table 2
Factor loadings from weighted structural equation model.

Leadership P-J Fit

The administration is supportive and encouraging 0.79

Staff members at this school are recognized for a job well done 0.81
My principal enforces school rules for student conduct and backs me up when I need it 0.76
The principal knows what kind of school he or she wants and has communicated it to the staff 0.69
The stress and disappointments in teaching at this school aren't really worth it 0.73
I don't have as much enthusiasm now as when I began teaching 0.76
I would certainly become a teacher if I had to go back and start over. 0.58
I think about staying home from school because I'm just too tired to go 0.50
I would leave as soon as possible for a higher paying job 0.60

All factor loadings are standardized.

Fit statistics CFA: c2 (20, N ¼ 3870) ¼ 214.544, p < 0.001; RMSEA ¼ 0.050; SRMR ¼ 0.030; CFI ¼ 0.985; and TLI ¼ 0.973.
Note: Fit statistics reported for unweighted structural equation model.

exponentiated values of the coefficients, which represent relative
risk ratios (RRR) relative to the base outcome of staying in the
current school. Relative risk ratios are interpreted as the ratio of
odds of the two outcomes that are being compared. For example, a
RRR of 1.0 indicates that the variable influences the odds of mobility
(moving schools or leaving teaching) at the same rate as it in-
fluences the base outcome (staying in the same school). Values of
the RRR less than 1.0 indicate that the variable reduces the odds of
mobility more than it reduces the odds of staying in the school.
Likewise, RRRs greater than 1.0 indicate the variable increases the
odds of mobility more than it increases the odds of staying in the
same school. Throughout what follows, we focus on the RRRs using
the value of 1.0 as the benchmark, and statistical significance is
always reported to indicate whether the RRR is statistically
different from 1.0. All models employ sampling weights to make the
results nationally representative and to account for the complex
sampling design of the SASS and TFS.
5. Results
The TFS survey included approximately 3800 respondents of
which approximately 3000 could be matched to their principal and
school surveys from the prior year.2 Because we expect the TFS
2
All sample sizes are rounded to the nearest 10 to comply with NCES data
reporting.
336 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339

Table 3 Table 4
Means and standard deviations of key variables. Multinomial logit models predicting teacher mobility based on fit measures.

Mean Standard Deviation Mover Leaver

Leadership 0.18 0.92 Leadership 0.746*** 1.052

P-J Factor 0.24 0.92 (0.003) (0.689)
Novice 7.9% 0.27 P-J fit 0.781** 0.551***
Female 76.0% 0.43 (0.020) (0.000)
Minority (teacher) 16.4% 0.37 Novice Teacher 1.584* 1.910
Married 68.6% 0.46 (0.078) (0.057)
Union Member 71.6% 0.43 Observations 3000 3000
Over 50 28.1% 0.45
***p < 0.01, **p < 0.05, *p < 0.1.
Master's degree 57.1% 0.50
p-values in parentheses.
Regular Certification 92.8% 0.26
Note: Analysis also includes controls for sex, race/ethnicity, marital status, union
Charter School 2.5% 0.16
membership, being over 50, master's degree, regular certification, race/ethnicity
Total Earnings $54,459 16,873.0
match with students, teachers, and principal; charter school, total earnings, number
Elementary School 49.9% 0.50
of LEP and IEP students, elementary school, special education teacher, math teacher,
Special Education 12.3% 0.33
science teacher, years at the current school, school size, schoolwide Title I status,
Math 8.1% 0.27
urbanicity, and schoolwide student and teacher demographics.
Science 7.5% 0.26
Years at School 4.3 4.26
Title I School 44.8% 0.50
Urban School 25.2% 0.44 While not the focus of our analyses, the regressions presented
Rural School 29.5% 0.46 above provide some important descriptive evidence about other
N 3000 predictors of teacher mobility holding constant the measures of
leadership and P-J fit. For instance, teachers in our sample in their
Note: Weighted means. Sample size is rounded to the nearest 10 to comply with
NCES reporting requirements. first three years of teaching (i.e., novice teachers) had higher odds
of leaving the profession and higher odds of moving from school to
another. Although both coefficients were rather large, neither was
sample to be representative of the nation as a whole, the weighted statistically significant at the conventional 0.05 level. The lack of
means should be roughly equivalent to what would be observed in statistical significance suggests that some of the attrition that is
the teaching population at large. However, because of non- seen among novice teachers is captured with the measures of fit
response at various survey levels, some differences may have and other controls that are included in the regression. This is
emerged. In our sample, approximately 8 percent of the teachers explored in greater detail below.
were novice teachers in 2011-12 (in their first three years of Other findings from the main results indicate that teachers with
teaching), roughly 2.5 percent of the sample taught in charter master's degrees were more likely to move to a new school
schools, and roughly half taught in elementary schools (defined as (RRR ¼ 1.861, p-value<0.01), but they were no more likely nor less
schools serving grades K-5) (See Table 3). Teachers had an average likely to exit the profession than teachers without a master's de-
of roughly 4 years of teaching experience at their schools. gree. Teachers with regular certification were less likely to change
schools than teachers who were alternatively certified
5.1. Principal leadership and teacher mobility (RRR ¼ 0.577, p-value<0.05). Teachers with higher salaries were
less likely to move to a new school (8.3 percent reduction in odds
As hypothesized, teachers who reported strong or high-quality per 10% increase, p-value<0.01) and not statistically different in
principal leadership were much less likely to leave their school of terms of their likelihood of exiting the profession. Teachers from
origin and move to another school (See Table 4). To illustrate, large schools (3.6 percent odds reduction per 100 students, p-
teachers who reported a level of principal leadership that is one value<0.05) and teachers from rural schools (RRR ¼ 0.699, p-
standard deviation greater than average had 25 percent lower odds value<0.1) were marginally significantly less likely to move schools,
of switching schools in the following school year (RRR ¼ 0.75, but neither school size nor urbanicity were related to teachers'
p < 0.01). However, teachers who reported more positive principal decisions to exit the profession. As expected due to retirements,
leadership were no more or less likely to leave the profession teachers over age 50 were much more likely to leave the profession
altogether; the coefficient was slightly greater than 1.0, but was not than younger teachers (RRR ¼ 4.107, p < 0.01) but were no more
close to being statistically significant. Taken together, these findings likely to switch schools.
confirm the hypothesis that teachers who reported stronger prin-
cipal leadership were less likely to switch from one school to 5.3. How teacher and school characteristics affect the role of
another than teachers who reported weaker principal leadership, leadership in teacher mobility
but principal leadership did not predict transitions out of the pro-
fession holding other factors (including P-J fit) constant. We predicted that a number of teacher and school characteris-
tics could influence the association between leadership and
5.2. Person-job fit and teacher mobility mobility. In what follows, we first describe whether average mea-
sures of leadership vary by teacher and school characteristics by
A teacher's measure of P-J fit is also predictive of mobility. examining the predictive power of such characteristics on mea-
Teachers who reported P-J fit of one standard deviation above sures of leadership. In this way, we can determine whether
average have 43 percent lower odds of leaving the profession (See particular teacher types or school types tend to report strong
Table 4, RRR ¼ 0.551, p-value<0.01). Contrary to our hypothesis, P-J leadership than others. We then test whether those characteristics
fit was also statistically related to the probability that a teacher influence the relationship between leadership and mobility by
changes schools; teachers who report greater P-J fit are less likely to interacting each characteristic with the leadership measure from
leave their school of origin and move to another school. The relative the primary specification.
risk ratio was 0.781 and was statistically significant at the five Several factors influenced measures of teachers’ perceptions of
percent level. leadership (See Table 5). Most notably, elementary teachers
 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339 337

Table 5 Table 6
Influence of school factors on principal leadership. Testing for factors that moderate principal leadership-teacher mobility association.

Leadership Mover Leaver

Elementary 0.178** Leadership * Elementary school 1.086 1.284

(0.013) (0.661) (0.218)
Novice 0.209* Leadership * Charter school 1.476 0.808
(0.058) (0.151) (0.420)
Female 0.150** Leadership * Rural school 1.094 1.233
(0.041) (0.630) (0.294)
Teacher minority 0.317* Leadership * Title I school 0.983 1.006
(0.078) (0.928) (0.971)
Married 0.016 Leadership * Novice teacher 0.700 0.784
(0.817) (0.178) (0.425)
Union member 0.118*
p-values in parentheses.
(0.074)
***p < 0.01, **p < 0.05, *p < 0.1.
Over 50 0.100
Note: Each interaction is from a separate multinomial logit regression. Each
(0.141)
regression includes indicators for principal leadership, P-J fit, and controls for sex,
Master's Degree 0.075
race/ethnicity, marital status, union membership, being over 50, master's degree,
(0.233)
regular certification, race/ethnicity match with students, teachers, and principal;
Regular Certification 0.164
charter school, total earnings, number of LEP and IEP students, elementary school,
(0.151)
special education teacher, math teacher, science teacher, years at the current school,
Teacher same race as teachers 0.088
school size, schoolwide Title I status, urbanicity, and schoolwide student and
(0.628)
teacher demographics.
Teacher-Principal race match 0.438
(0.294)
Teacher-Principal race match 0.071
(0.436) reported stronger leadership (0.28 units greater) than that reported
Charter 0.071 by other teachers while none of the other subject areas were sta-
(0.453) tistically different from zero.
Total Earnings (log) 0.098
Given these differences in perceptions of leadership, we exam-
(0.416)
LEP students taught 0.000 ined whether some of these teacher and school characteristics
(0.822) moderated the influence of principal leadership on teacher
IEP students taught 0.001 mobility (See Table 6). Elementary and secondary schools differ in
(0.317)
structure and therefore may differ in the relationships between
Special Education 0.284***
(0.001) teachers and subsequently how closely leadership and fit are
Math 0.011 related to teacher mobility decisions. To test this, we interacted
(0.923) measures of leadership with the school level taught. Contrary to
Science 0.193 our hypotheses, the coefficients for both outcomes were positive,
(0.122)
which suggests that if anything elementary teachers experience a
Years at School 0.008
(0.213) less sensitive connection between leadership and mobility than
Female Principal 0.106* secondary teachers. However, both coefficients were modest and
(0.083) neither was statistically significant at the 0.05 level.
School size 0.002
Compared to traditional public schools, charter schools also vary
(0.692)
Title I 0.100
in their structure in ways that may influence how leadership is
(0.127) related to mobility. Regressions that interacted charter school in-
Urban 0.053 dicators with leadership, though, found no statistically significant
(0.497) relationship between the effect of leadership in charter schools
Rural 0.143**
versus leadership's effect in traditional public schools. This suggests
(0.033)
% Minority teachers 0.003 that leadership is no more or less important for charter school
(0.236) teachers than it is for traditional public school teachers as it relates
% Minority students 0.001 to mobility decisions.
(0.712)
School urbanicity influenced teacher mobility patterns in our
0.178**
Observations 3000
primary specifications. Urbanicity could also influence the rela-
R-squared 0.054 tionship between fit and mobility since in less-densely populated
rural areas, for example, it might be much more difficult to transfer
p-values in parentheses.
***p < 0.01, **p < 0.05, *p < 0.1. to other schools than it is in suburban or urban settings where
schools are more concentrated. When we interacted rural school
location with leadership, we found that urbanicity did not have
reported, on average, stronger leadership (i.e., 0.17 units greater) significant explanatory power in predicting teacher mobility
than non-elementary teachers and novice teachers reported lead- related to leadership. This suggests that the relationship between
ership that was stronger than that reported by more experienced leadership and mobility is not different for rural teachers compared
teachers, although this difference was only marginally significant. to urban or suburban teachers. This is a somewhat surprising result
Female teachers reported weaker measures of leadership than their given assumed differences in the availability of opportunities to
male colleagues, but minority teachers reported stronger measures move.
than White teachers. Teachers who were members of a teacher The demographics of students could influence turnover, per-
union reported leadership that was weaker (i.e., 0.12 units lower) ceptions of leadership, and feelings of fit. As reported above,
than that reported by non-union teachers, and teachers who were teachers from Title I schools reported lower leadership than
over 50 reported leadership that was weaker (0.12 units lower) teachers in non-Title I schools. When interacted with leadership,
than reported by younger teachers. Special education teachers
338 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339
however, being in a Title I school did not have a significant influ-
ence on the relationship between leadership and mobility.
As a final exploration, we examined whether the relationship
between leadership and mobility was different for novice teachers
than it was for non-novice teachers. To test whether novice teacher
status moderated the effects of leadership on mobility, we inter-
acted such status with leadership. Contrary to our hypothesis, the
interaction between leadership and novice teacher status was not
statistically significant, which suggests that the relationship be-
tween leadership and mobility is no different for novice teachers
than it is for more experienced teachers.
6. Discussion
Due to its potential impact on students and school climate, a
substantial body of prior research has examined the determinants
of teacher mobility (Borman & Dowling, 2008; Guarino et al., 2006).
While existing studies of teacher retention have primarily
attempted to isolate economic and, to a lesser degree, organiza-
tional factors that predict teacher turnover, this analysis extended
the teacher retention research base by employing a person-
environment fit theoretical framework, frequently used in studies
of turnover in other professions, but infrequently seen in studies
focusing on teachers. Specifically, we explored how principal
leadership and person-job (P-J) fit were associated with teacher
retention.
Through multiple models and specifications, the results confirm
that principal leadership and P-J fit predict retention in one's school
and retention in the teaching profession, respectively. Since per-
ceptions of principal leadership reflect a teacher's sense of fit with
their school and not their general satisfaction with teaching, it
makes intuitive sense that strong leadership is negatively corre-
lated with a teacher's decision to move from school to another.
Likewise, P-J fit is a measure of how well the teacher fits with the
larger teaching profession. These findings are consistent with other
research from within and outside the U.S. on principal leadership
and teacher retention (Allensworth et al., 2009; Boyd et al., 2011b;
Hulpia et al., 2011; Ladd, 2011; Skaalvik & Skaalvik, 2011).
In terms of implications for leadership practice, these results
strongly suggest that certain principal leadership behaviors can
directly impact teacher retention in their current schools. These
behaviors include principals communicating a vision for the school
to their staff and working to achieve that vision, being supportive to
teachers with regard to instruction and other issues, recognizing
exemplary teaching performance, and enforcing rules related to
student behavior and discipline. These findings also have implica-
tions for school districts and charter management organizations
(CMOs). By focusing on such leadership behaviors in principal
evaluation and professional development activities, districts and
CMOs can indirectly promote teacher retention. Finally, principal
preparation programs can support teacher retention by helping
school leaders acquire knowledge and skills related to these lead-
ership behaviors during formal preparation.
In our analyses, principal leadership did not predict retention in
the profession. Also, consistent with our expectation, we found no
difference with regard to the relationship between leadership and
retention in one's school when comparing teachers in rural schools
versus those in urban and suburban schools. This suggests that
leadership behaviors have a consistent impact on teachers' expe-
riences and commitment to their schools across many different
school locations. Contrary to our hypotheses, we found no statis-
tically significant differences with regard to the relationship be-
tween school leadership and retention in one's school when
comparing novice and experienced teachers, elementary and sec-
ondary teachers, teachers in charter schools and those in traditional
public schools, or teachers in Title I schools and those in non-Title I
schools. The robustness of these findings suggests that the strength
of the association between leadership and retention is remarkably
consistent across a wide variety of teaching contexts, including
international teaching contexts.
P-J fit is a measure of how well the teacher fits with the larger
teaching profession. As expected, we found a strong association
between high P-J fit and a reduced likelihood that teachers would
leave the profession. This finding is consistent with findings from I-
O psychology research (Kristof-Brown et al., 2005) and research in
K-12 education (Kelly & Northrop, 2015; Tickle et al., 2011). In
addition, we were surprised to find that P-J fit was also associated
with retention in one's school. It seems that teachers who perceive
a strong fit between their abilities or needs and the demands of the
teaching profession are likely to be committed to their schools as
well. Taken together, these results indicate that teachers will be
more likely to continue in their schools and in the teaching pro-
fession when they feel that their needs and abilities match the
requirements of the profession.
In terms of implications for pre-service teacher education pro-
grams, the results suggest that preparation programs should pro-
vide teaching candidates with extensive opportunities to simulate
the challenges and lived experiences of practicing teachers through
student teaching placements, other field experiences, and simu-
lated laboratory settings. With regard to international settings,
these findings indicate that efforts to promote teachers’ percep-
tions of P-J fit are likely to promote retention across different na-
tional contexts.
6.1. Limitations
While these results offer useful evidence of relationships be-
tween principal leadership, P-J fit, and teacher retention, there are
some limitations to our analysis that warrant discussion. First, this
study did not include any measures of student achievement or
teacher effectiveness. Recent studies have focused on differential
retention and have found that teachers who are deemed to be
“more effective” than their peers are more likely to be retained,
while those who are “least effective” are more likely to leave
(Goldhaber, et al., 2011; Boyd et al., 2011b). An important step for
future research will be to build on these studies by integrating
teacher effectiveness measures into analysis of leadership, fit, and
retention and examining how fit interacts with teacher effective-
ness to shape teacher retention, migration, and attrition.
Second, the SASS and TFS data make it impossible to distinguish
between permanent leavers and those who “stop out” to pursue
other opportunities (e.g., graduate school, caring for a relative,
having a baby) for a period of time before returning to teach in the
classroom. In this analysis, these individuals can only be considered
“leavers.” Ingersoll and May (2012) note that temporary attrition
leads to school-level staffing challenges that are similar to complete
exit, but do not quantify the percentage of teachers who leave a
school but eventually re-enter the classroom. Estimates of the
percentage of teachers who “stop out” are scarce, but Provasnik and
Dorfman (2005) estimated that about 4 percent of the teachers who
entered a new school in the 1999e2000 academic year were
returning to the classroom after taking time off, which was rela-
tively consistent with estimates from the prior 10 years. Given that
“stopping out” is typically related to life circumstances such as child
rearing, it is unlikely that this form of temporary attrition would be
significantly related to the types of leadership or fit used in this
analysis.
Third, the data for this study comes from a nationally repre-
sentative sample of teachers in the U.S. In future research, it would
be useful to include principals and teachers from multiple countries
 D. Player et al. / Teaching and Teacher Education 67 (2017) 330e339
in order to examine whether principal leadership and P-J fit have
similar associations with teacher retention in other national
contexts.
Moving forward, principal leadership, P-J fit, and teacher
retention warrant continued attention in educational policy and
research, particularly when one aim is to increase retention of
highly effective teachers while minimizing costly efforts to retain
lower performers who are unlikely to improve. Even the most
talented teachers will be unable to reach their full potential if their
teaching positions are not a good fit. Using recruitment and se-
lection policies to match teachers with the environments in which
they are most likely to be successful is a promising strategy for
improving both retention and student achievement. As policy
continues to emphasize using teacher effectiveness as a component
of teacher and school evaluation, fit with one's school and with the
teaching profession should also remain a focus of policymakers,
practitioners, and researchers.
References
Allensworth, E., Ponisciak, S., & Mazzeo, C. (2009). The schools teachers leave:
Teacher mobility in Chicago public schools. Chicago: Consortium on Chicago
School Research at the University of Chicago Urban Education Institute.
Borman, G., & Dowling, N. M. (2008). Teacher attrition and retention: A meta-
analytic and narrative review of the research. Review of Educational Research,
78(3), 367e409.
Boyd, D., Grossman, P., Ing, M., Lankford, H., Loeb, S., & Wyckoff, J. (2011). The in-
fluence of school administrators on teacher retention decisions. American
Educational Research Journal, 48(2), 303e333.
Boyd, D., Lankford, H., Loeb, S., Ronfeldt, M., & Wyckoff, J. (2011). The role of teacher
quality in retention and hiring: Using applications to transfer to uncover
preferences of teachers and schools. Journal of Policy Analysis and Management,
30(1), 88e110.
Cable, D., & Judge, T. (1996). Person-organization fit, job choice decisions, and
organizational entry. Organizational Behavior and Human Decision Processes,
67(3), 294e311.
Caldwell, D., & O'Reilly, C. (1990). Measuring person-job fit with a profile-
comparison process. Journal of Applied Psychology, 75(6), 648e657.
Chan, W. Y., Lau, S., Nie, Y., Lim, S., & Hogan, D. (2008). Organizational and personal
predictors of teacher commitment: The mediating role of teacher efficacy and
identification with school. American Educational Research Journal, 45(3),
597e630.
Chatman, J. (1989). Improving interactional organizational research: A model of
person-organization fit. Academy of Management Review, 14(3), 333e349.
Clotfelter, C. T., Ladd, H. F., Vigdor, J. L., & Diaz, R. A. (2004). Do school accountability
systems make more difficult for low-performing schools to attract and retain
high-quality teachers? Journal of Policy Analysis and Management, 23(2),
251e271.
Coldwell, M. (2016). Career orientations and career cultures: Individual and
organisational approaches to beginning teachers' careers. Teachers and Teach-
ing: Theory and Practice, 22(5), 610e624.
Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis (2nd ed.). Hillsdale,
NJ: Erlbaum.
De Nobile, J. J., & McCormick, J. (2008). Organizational communication and job
satisfaction in Australian Catholic primary schools. Educational Management,
Administration, & Leadership, 36(1), 101e122.
Edwards, J. (1991). Person-job fit: A conceptual integration, literature review, and
methodological critique. In C. L. Cooper, & I. T. Robertson (Eds.), International
review of industrial organizational psychology (Vol. 6, pp. 283e357). Chichester,
England: Wiley.
Edwards, J., & Shipp, A. (2007). The relationship between person-environment fit
and outcomes: An integrative theoretical framework. In C. Ostroff, & T. Judge
(Eds.), Perspectives on organizational fit. New York: Taylor and Francis.
Goldhaber, D., Gross, B., & Player, D. (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), 57e87.
Goldring, R., Taie, S., & Riddles, M. (2014). Teacher attrition and mobility: Results from
the 2012-13 teacher follow-up survey. NCES 2014-077). U.S. Department of Ed-
ucation. Washington, DC: National Center for Education Statistics.
339
Guarino, C., Santibanez, L., & Daley, G. (2006). Teacher recruitment and retention: A
review of the recent empirical literature. Review of Educational Research, 76(2),
173e208.
Hanushek, E., Kain, J., & Rivkin, S. (2004). Why public schools lose teachers. Journal
of Human Resources, 39, 326e354.
Hoffman, B., & Woehr, D. (2006). A quantitative review of the relationship between
person-organization fit and behavioral outcomes. Journal of Vocational Behavior,
68(3), 389e399.
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling: A Multidisciplinary Journal, 6(1), 1e55.
Hulpia, H., Devos, G., & Van Keer, H. (2011). The relation between school leadership
from a distributed perspective and teachers' organizational commitment:
Examining the source of the leadership function. Educational Administration
Quarterly, 47(5), 728e771.
Ingersoll, R. (2001). Teacher turnover and teacher shortages: An organizational
analysis. American Educational Research Journal, 38(3), 499e534.
Ingersoll, R. M., & May, H. (2012). The magnitude, destinations, and determinants of
mathematics and science teacher turnover. Educational Evaluation and Policy
Analysis, 34(4), 435e464.
Keigher, A. (2010). Teacher attrition and mobility: Results from the 2008e09 teacher
follow-up survey (NCES 2010-353). U.S. Department of Education. Washington,
DC: National Center for Education Statistics.
Kelly, S., & Northrop, L. (2015). Early career outcomes for the “Best and the
Brightest”: Selectivity, satisfaction, and attrition in the Beginning Teacher
Longitudinal Survey. American Educational Research Journal, 52(4), 624e656.
Kristof-Brown, A., Jansen, K., & Colbert, E. (2002). A policy-capturing study of the
simultaneous effects of fit with jobs, groups, and organizations. Journal of
Applied Psychology, 87(5), 985e993.
Kristof-Brown, A., Zimmerman, R., & Johnson, E. (2005). Consequences of in-
dividuals' fit at work: A meta-analysis of person-job, person-organization,
person-group, and person-supervisor fit. Personnel Psychology, 58(2), 281e342.
Ladd, H. F. (2011). Teachers' perceptions of their working conditions: How predic-
tive of planned and actual teacher movement? Educational Evaluation and Policy
Analysis, 33(2), 235e261.
Lankford, H., Loeb, S., & Wyckoff, J. (2002). Teacher sorting and the plight of urban
schools: A descriptive analysis. Educational Evaluation and Policy Analysis, 24(1),
37e62.
Lauver, K., & Kristof-Brown, A. (2001). Distinguishing between employees' per-
ceptions of person-job and person-organization fit. Journal of Vocational
Behavior, 59(3), 454e470.
Muchinsky, P., & Monahan, C. (1987). What is person-environment congruence?
supplementary vs. complementary models of fit. Journal of Vocational Behavior,
31(3), 268e277.
Piasentin, K., & Chapman, D. (2006). Subjective person-organization fit: Bridging
the gap between conceptualization and measurement. Journal of Vocational
Behavior, 69(2), 202e221.
Pogodzinski, B., Youngs, P., & Frank, K. (2013). Collegial climate and novice teachers'
intent to remain in teaching. American Journal of Education, 120(1), 27e54.
Provasnik, S., & Dorfman, S. (2005). Mobility in the teacher workforce: Findings from
the condition of education 2005. NCES 2005-114. Washington, DC: National
Center for Education Statistics.
Ronfeldt, M., Loeb, S., & Wyckoff, J. (2013). How teacher turnover harms student
achievement. American Educational Research Journal, 50(1), 4e36.
Ryan, A. M., & Schmit, M. (1996). An assessment of organizational climate and P-E
fit: A tool for organizational change. The International Journal of Organizational
Analysis, 4(1), 75e95.
Scafidi, B., Sjoquist, D. L., & Stinebrickner, T. R. (2007). Race, poverty, and teacher
mobility. Economics of Education Review, 26(2), 145e159.
Skaalvik, E. M., & Skaalvik, S. (2011). Teacher job satisfaction and motivation to leave
the teaching profession: Relations with school context, feeling of belonging,
and emotional exhaustion. Teaching and Teacher Education, 27(6), 1029e1038.
Smith, T. M., & Ingersoll, R. M. (2004). Reducing teacher turnover: What are the
components of effective induction? American Educational Research Journal,
41(3), 681e714.
Stuit, D. A., & Smith, T. M. (2012). Explaining the gap in charter and traditional
public school teacher turnover rates. Economics of Education Review, 31(2),
268e279.
Tickle, B. R., Chang, M., & Kim, S. (2011). Administrative support and its mediating
effect on U.S. public school teachers. Teaching and Teacher Education, 27(2),
342e349.
Van Maele, D., & Van Houtte, M. (2012). The role of teacher and faculty trust in
forming teachers' job satisfaction: Do years of experience make a difference?
Teaching and Teacher Education, 28(6), 879e889.