Deploy your best talent to the classrooms where they're needed the most and, abracadabra! kids learn more! 

So concludes a new report from Mathematica out this week.  

In seven unnamed districts, really great teachers were given $20,000 to transfer to a lower-achieving school in the same district. Test scores went up considerably, 4 to 10 percentile points. Nothing magical about those solid figures, but underneath the strong results are some important lessons.

First, while giving a teacher $20,000 to transfer within a district may seem like an awful lot of money, it's quite a bit cheaper than the much more politically popular approach of reducing class size. By Mathematica's estimates, that strategy would have achieved a similar result at a much higher cost -- $27,000.

What's also encouraging here is the retention. Almost all of the teachers in the initiative honored their two-year commitment. Afterward, about two-thirds elected to stay a third year. (It's too soon to report 4th year retention.)

A few caveats. Those positive results didn't show up at the middle school level where the experiment was also tried -- for reasons which aren't clear. Also, the report doesn't describe what happened to the test scores of the schools in which the teachers had formerly taught.  For all we know, the district traded better test scores in one school for worse in another, not a good result (although one can certainly argue that equity goals were paramount). 

Perhaps the biggest stumbling block here, and one that we have observed for similar efforts such as Charlotte-Mecklenberg's, is that it's not easy to get superstar teachers to consider a move. Of the 1,500 teachers who qualified to participate in this program, only a third showed up to the initial information session, and ultimately only 81 teachers across all seven districts took on the challenge, leaving a number of the potential transfer vacancies unfilled.