Imagine a test where the only grades a student could earn were A or A-. Envision an update to Yelp that allows customers to rate restaurants using only 4 or 5 stars. Picture a panel of judges at an ice-skating competition where the only scorecards say 9 or 10.

That's what's going on with teacher evaluations, despite a long and brutish effort to get states and districts to do otherwise. Most teachers continue to receive high ratings and there's very little evidence that anyone is below average. (Our recent report, Running in Place, describes how state evaluation systems enable teachers to continue earning top ratings regardless of how their students fare.)

A recent study from Jason Grissom of Vanderbilt and Susanna Loeb of UCLA provides evidence that principals just aren't getting it.

Grissom and Loeb investigated how principals evaluate teachers formally versus how they view them privately. They selected roughly 100 principals from Miami-Dade Public Schools, asking them in interviews to evaluate four of their teachers. The rubric they used mirrored the district's formal evaluation framework.

Even in the informal setting, principals rated most of their teachers on the high end, but were still tougher than they were on the formal evaluation. With those, they rated fewer than one percent of their teachers as "ineffective" or "needs improvement" on any standard except professionalism. On the informal evaluation, they rated roughly 20 percent of teachers accordingly.

There is a silver lining. Yes, principals are reluctant to speak the truth, but the two sets of ratings did correlate. The teachers rated lower on the informal ratings were the lower rated teachers on the formal ratings. Better yet, both ratings correlated with teacher value-added scores, meaning that principals know good teaching from not-so-good teaching--they just don't call it out.

On a positive note, a recent analysis of a principal training program in Houston suggests that professional development of principals may help address these types of managerial challenges. In a school-level randomized field experiment in Houston, Roland Fryer (Harvard University) found that attention to increasing principal management skills led to statistically significant increases in teacher performance across the board, as measured by their student learning gains in just the first year. The training involved 300 hours across two calendar years with a focus on instructional planning, data-driven instruction, and, importantly, observation and coaching. 

This benefit for teachers and students, coupled with the relatively low marginal cost of providing the professional development, makes principal management training worth considering.