The most pressing conclusions come from the survey results about school
administration. Survey questions about school administration included questions about
behavior toward staff, as well as questions about administrator practices such as data c
ollection and resource gathering. This research suggests that teachers who have less-positive
perceptions of their school administrators are more likely to transfer schools or leave teaching
in New York City Public Schools altogether (2011, p. 323). The authors conclude that
administrative support accounts for large differences in attrition rates across schools (2011, p.
323). In addition to measuring the influence of administrators on a teacher’s decision to stay
or leave, the researchers suggest that these results offer promising avenues for policy changes.
While student backgrounds cannot be changed, policies to improve school leadership can be
one avenue that increases teacher retention (2011, p. 329).
Teacher Movement within New York City Public Schools
An additional study conducted by Boyd, Lankford, Loeb, Ronfeldt, and Wyckoff
examines which teachers applied to transfer between schools and were successful across two
consecutive school years. The authors state “two substantive questions guide [their] study:
Which teachers are more likely to request transfers? What kinds of teachers do schools choose
to hire?” (2010) To answer these questions, the authors use both descriptive statistics and
logistic regression.
Boyd, Lankford, Loeb, Ronfeldt, and Wyckoff found that “white teachers [tend to] teach
in schools with the fewest poor and lowest-achieving students” (2010, p. 94) and that these
same schools tend to have the most high-quality teachers (2010, p. 96). The measures of quality
used for this study are different than previous studies summarized here in that teacher
quality is measured primarily through pre-certification factors such as the prestige level of
their undergraduate college and their scores on the New York state certification exam (2010, p.
93). This suggests that teachers who have competitive pre-service qualifications seek to avoid
specific schools by using the transfer process; the authors state, “the kinds of schools that
receive the fewest applicants are in many ways similar to the kinds of schools that teachers
want to leave” (2010, p. 97).
Altogether, the two most predictive factors for likelihood to transfer were teacher race
and number of years teaching (2010, p. 99, 101). White teachers with fewer years of
experience were more likely to request to transfer than their colleagues of other races from
the same schools and more likely to request to transfer than their white colleagues who had
been teaching for longer (2010, p. 101). The authors stress that this data cannot fully express
the range of reasons why these patterns emerge (2010, p. 102), and note that even though white
teachers request to transfer more frequently, “black and Hispanic teachers are significantly
more likely than white teachers to get hired” (2010, p. 107). This conclusion seems to suggest
that further explorations of the dynamics of race could be uncovered with future research.
Value-Added Modeling: Analyzing the Relationship Between Measures of Quality and Retention
Winters and Cohen (2013) analyzed teacher evaluation and dismissal policies across
18 states and Washington, D.C. They first identified which of the 19 locales used value-added
modeling for teacher evaluations. Value-added modeling is “measured as a contribution to
student academic growth,” and is typically quantified using state standardized assessments
(2013, p. 636). In Florida, for example, a teacher’s value-added modeling score is calculated
using student growth or proficiency scores on a test such as the Florida Standards Assessment
or an Advanced Placement exam. In all but 6 of the 19 locales examined, teacher performance
evaluations were based “primarily on student growth scores” (2013, p. 636).
8