Virtually all discussions of teacher turnover focuses on teachers leaving schools and/or the profession. However, a recent working paper by Allison Atteberry, Susanna Loeb and James Wyckoff, which was presented at this month’s CALDER conference, reaches a very interesting conclusion using data from New York City: There is actually more movement within NYC schools than between them.*
Specifically, the authors show that, during the years for which they had data (1997-2002 and 2004-2010), over 50 percent of teachers in any given year exhibited some form of movement (including leaving the profession or switching schools), but two-thirds of these moves were within schools – i.e., teachers changing grades or subjects. Moreover, they find that these within-school moves, like those between-schools/professions, appear to have a negative impact on testing outcomes, one which is very modest but statistically discernible in both math and reading.
There are a couple of interesting points related to these main findings.
The first is the sheer extent of within-school churn. It may be the case that NYC is somewhat of an outlier here, but the fact that one-third of teachers in the city stay in the same school but change grades/subjects in any given year is rather surprising (at least to me). And this analysis suggests that it may also have consequences for teacher and student performance.
(Side note: Needless to say, these within-school moves, both in terms of why they happen and what they mean, vary between elementary, middle and high schools.**)
Another fascinating aspect of this study is the obvious question of why all this movement occurs. No doubt some of it is due to “natural causes." For example, schools with high attrition/mobility rates might have to shuffle people around every year to fill the spots of those who left (this paper does find a correlation between within-school churn and turnover rates).
As the authors point out, it’s also quite possible that school administrators are employing accountability strategies, such as trying to get their best teachers into tested grades and subjects so as to boost their testing results (see here). Note, however, that they find little evidence of this kind of strategy (and that overall movement rates are relatively stable across years).
There also appears to be quite a bit of variation between schools in terms of how much churn they exhibit. The rates vary from virtually no within-school churn to around 50-60 percent. This suggests that there may be something about school leadership or culture that generates this tremendous fluidity in assignments within schools.
Finally, one huge factor, at least in NYC, may be teacher preferences. The guidelines surrounding assignment in the teacher contract (see Article Seven) vary a bit between elementary and secondary schools, but the general idea is that teachers express their preferences for assignments ahead of time, administrators are supposed to grant those preferences to the degree possible/advisable. It seems, therefore, that within-school assignments reflect an interplay of teacher/administrator preferences as well as logistics, such as the need to fill vacancies due to attrition.
In any case, the bottom line here is that there seems to be an annual shuffle of assignments within NYC schools, and while some (perhaps most) of it is necessary and/or beneficial, much of it is “voluntary," and seems to have a modest negative impact overall on test-based outcomes. Going forward, it will be very interesting to see whether movement is as high in other districts (and whether it's associated with test-based outcomes).
- Matt Di Carlo
* The link directs you to the presentation slides. The full paper will be available soon from CALDER.
** NYC high school teachers, for example, rarely teach outside of their license area (i.e., switch subjects), but may teach different grades within that subject in any given year (e.g., three classes of ninth grade English and two classes of eleventh grade English). If some of those grades were to change year-to-year (e.g., instead of two classes of eleventh grade English she teaches two of tenth grade), that is a rather less “drastic” move than an elementary school teacher switching grades. The authors disaggregate their results by school level, and find that overall movement (attrition/mobility plus within-school switching) is indeed higher in middle and high schools, but still rather high in elementary schools.