Lots of studies have tried to pinpoint the characteristics that make for an effective teacher. Not so many have tried to do the same for principals. Is a more experienced principal better at the job than, say, a newbie? What about experience as an assistant principal or a teacher? Does a degree from an Ivy League college signal that a principal will really know how to run a school? The research base on all of these questions is paper thin.
A new working paper attempts to plug in some of the gaps in that knowledge base. Looking at 20 years of data on hundreds of principals in the New York City public school system, researchers Damon Clark, Paco Martorell, and Jonah Rockoff try to determine which qualities of principals seem to be most closely linked to good outcomes for their schools, such as gains in student achievement, fewer teacher and student absences, or lower rates of teacher turnover.
It’s true in New York, as in some other places, that principals are not randomly assigned to schools. More-experienced principals, for instance, will tend to move to higher-performing schools over time, if given the choice, or, conversely, they may be called in to head the schools with the worst achievement track records. But the researchers control for socioeconomic and achievement differences across schools by comparing principals to their successors and predecessors in the same buildings. This was particularly easy to do in New York City because of a 1991 retirement incentive that induced a quarter of the city’s 1,000 principals to retire that year.
The researchers’ clearest finding is one that you might expect: Principals get better at their jobs with every year of experience, and particularly so in the first few years on the job. If that sounds familiar, it may be because there’s a similar, oft-quoted finding about teachers. They get better with experience, but only up until the third year on the job.
Having graduated from a highly selective university, on the other hand, or spending time as a classroom teacher seemed to be less important for the principals in this analysis. Experience as an assistant principal also seemed to lead to better school outcomes, the researchers found, but only for inexperienced principals.
Findings were also mixed on whether principals who had been through principal training programs were any more effective than their untrained counterparts. Performance improved at schools headed by principals who took part in the Cahn Fellows Program, which is run by Teachers College, Columbia University. When a school was assigned a principal from the Aspiring Principals Program, however, school performance did not improve---at least not in the short run. But the Cahn Fellows are experienced principals handpicked from the city’s highest-achieving schools. The APP principals are newbies by definition. The study also found that they tend to be assigned to schools that already seem to be on a downward trajectory, which makes it difficult to draw any definitive conclusions on the relative success of those professional-development programs.
The researchers said their finding on the value of experience raises some important policy implications. “First, it alerts district administrators to the potential costs of having experienced principals leave their jobs or equivalently, informs them of the benefits associated with retaining experienced principals,” the researchers write. “Second, it alerts district administrators to the distributional consequences that follow from higher rates of turnover in disadvantaged schools.”
As careful as these researchers were, however, this study is not a randomized experiment, which means that you really can’t say any of the characteristics identified cause better school performance. To strengthen their analysis, they plan to take advantage of a natural experiment they found in the data. Under the terms of that 1991 retirement buyout, principals had to be a certain age or have accrued certain levels of experience to qualify. That allows the researchers to go back into the data and compare principals who just missed qualifying for the incentive with those who just made it. The method is what researchers call a"regression discontinuity design” and experts say it’s the next best thing to a pure experiment for figuring out questions of causation. So stay tuned.
This paper is among a batch of new papers posted this week by the Center for Analysis of Longitudinal Date in Education Research or CALDER, a federal research center operated by the Urban Institute. Check them all out here. Follow this link to read EdWeek’s story on the conference the institute held last month to showcase all those papers.