Teacher Training
One of the key components of traditional teacher preparation programs is field experience. The
data analyzed in Ronfeldt’s 2012 study came from surveying teachers teaching in New York City Public
Schools who also experienced teacher preparation experiences, or field experience placements, in New
York City Public Schools. According to Ronfeldt, there are two opposing theories about what types of
schools are appropriate for field placement of new teachers. One theory suggests that field placement
should occur in schools that are considered to be easy-to-staff so that teachers will see examples of
ideal teaching practices. The opposing theory is that teachers should train in difficult-to-staff schools
so that they have more realistic expectations of their possible teaching placement. In these two oppos-
ing models, the types of schools described as easy-to-staff schools tend to have white students from
affluent families. These students have fewer academic barriers, whereas difficult-to-staff schools are
described as having students of color from varying demographic groups and students who are of a
lower socioeconomic status (2012, p. 6). However, Ronfeldt challenges the notion that difficult-to-staff
schools should be defined solely on student characteristics (2012, p. 8). This research suggests that
initiating an analysis of school-level teacher turnover is a critical measure because student
characteristics do not provide a comprehensive picture of the practical reasons why teachers leave
schools.
The original findings that Ronfeldt describes include a finding that suggests that teachers who
train in schools with a higher “stay-ratio” are less likely to leave New York City schools within their
first five years of teaching (2012, p. 13). The stay-ratio Ronfeldt developed measures how likely teachers
are to leave a given school in any given year (2012, p. 8). This research shows that student backgrounds
are not a significant predictor of teacher turnover at a particular school, but that other measurable
pieces of school culture are significant predictors of teacher turnover. This research is important
because it shifts the focus of teacher placement to school culture with strong support systems and
demonstrated histories of implementing best practices, regardless of student backgrounds. The two
most important takeaways from this research are: teachers should be placed in easy-to-staff schools
for their training, and easy-to-staff schools cannot be defined solely based on student characteristics
(2012, p. 22). An important policy recommendation made by the authors of this study is to routinely
calculate the stay-ratio for schools within a given school district and to use that information as a
consideration when determining teacher-training placements.
Teacher Evaluation and Administrative Support
While a stay-ratio is a measure for how many teachers stay in a given school or district in
a given year, it does not answer the question of why teachers choose to stay rather than leave. The
following studies examine different aspects of schools that may have some bearing on an individual
teacher’s decision to stay in the profession or leave. The first study examines the relationship between
perception of administrator quality and a teacher’s decision to stay or leave. The second study analyzes
the relationship between financial incentives tied to student growth and a teacher’s decision to stay or
leave.
The Role of Administrators in Teacher Retention
The study by Boyd, et. al. examined the relationships between first year teacher
perceptions of school contextual factors and what, if any, relationship exists between
those factors and teacher attrition. (2011, p. 311) One of the stated goals of the study was
to build on prior research that assessed the effects of working conditions on teacher
retention (2011, p. 309). This research also studied first year teachers in New York City
Public Schools and used administrative data from New York City Public Schools to
track retention behaviors across the school district (2011, p. 310). The survey was broken
into several areas and respondents used a Likert scale of 1-5 to respond to each survey
item.
7