TQB: Teacher Quality Bulletin

The numbers don’t lie-but they may bend the truth a little

See all posts

Here's a question which (as far as we know) no one has asked before: are new teachers who graduate from more elite colleges more likely to quit?

Sean Kelly and Laura Northrop of the University of Pittsburgh ask and answer this question, reporting that teachers who attend highly selective institutions are more likely to leave the profession than their peers. 

However, we're crying foul. 

State Kelly and Northrop: graduates of highly selective institutions have an estimated "85 percent greater likelihood of leaving the profession than less selective graduates in the first three years of teaching" (p. 25). Wow. That's bad, right?

Before districts start turning away any applicants from the Ivy League, read the fine print. The researchers used data from a sample of teachers (via the federal government's Beginning Teacher Longitudinal Survey) to estimate that across all teachers, about 15 percent of graduates from highly selective colleges are likely to leave the profession within three years, compared to eight percent of graduates from less selective institutions. True, that roughly seven percentage point difference (15 percent minus eight percent) does represent an 85 percent increase, as Kelly and Northrop claim, but, also true, it's a distinction without a difference. 

First, a little about statistics. An increase from 10 to 15 percent or from 90 to 95 percent is, in both instances, a five percentage point increase. However, in the former, it's a 50 percent increase; in the latter, it's only a 5.5 percent increase. Clearly, choosing to use percents (as Kelly and Northrop do) rather than percentage points is misleading…while still technically accurate. 

Second, and more to the point, while the different attrition rates were not negligible, they also weren't statistically significant. Kelly and Northrop attribute the lack of statistical significance to the small sample size (160 graduates from highly selective colleges, and 1,350 graduates from less selective colleges), but that doesn't tamp down their enthusiasm for repeatedly referring to the difference in attrition rate—even though it was statistically indistinguishable from zero. 

Less time was spent on findings that were statistically significant. For example, we also learn that teachers who started their teaching career later in life were found to be more likely to leave the profession within three years. Teachers who were earning relatively higher salaries were found to be less likely to leave.