A recent working paper by Michael Hansen looks at formerly low-performing schools that improved to try to figure out how they accomplished this feat.

In his analysis of six years of Florida and North Carolina data, Hansen examines schools that the study labels turnaround schools. This term is used informally and is not meant to convey a specific improvement model. Instead, Hansen chose schools that were originally low-performing (low in status and growth) and showed progress (an increase in status and well-above-average growth).

Hansen focuses on the teachers and principals in these schools, hypothesizing that at least one of two things happened:

1.  Teachers (and principals) who remained ("stable" teachers) showed improvement, or

2.  Low-performing teachers (and principals) left the school and were replaced by higher-performers.

In fact, both happened. The biggest effect came from the stable teachers--they showed substantial improvement. In addition, the incoming teachers outperformed those that left, so these new teachers also contributed to the schools' improvement. 

Statistically, the exit of lower-performing teachers alone did not improve school quality: low-performing teachers left both the schools that improved and those that didn't  

Of course, this retrospective study cannot determine what caused the turnaround. The essential questions of why and how those stable teachers became more effective aren't answered. There's so much we still don't know about what staff supports, hiring policies or other changes these successful schools implemented and whether these changes can be replicated to produce the same results in other struggling schools. But even without the full back-story, this study does suggest that schools should focus their efforts on helping the teachers who remain to improve and on making good new hires, rather than just trying to weed out the low performers.