Breaking new research ground, Cornell economist C. Kirabo Jackson suggests that fully a quarter of a teacher's effectiveness may be actually attributable to how well a teacher fits in a school.
Jackson uses 11 years worth of teacher assignment data from North Carolina to find that teachers tend over the years to gravitate where they best fit. High performing teachers gravitate to schools which share their teaching philosophies and work cultures. Lower performing teachers gravitate to schools with low expectations.
Jackson finds that more experienced teachers (those with 10 to 20 years of experience) are likely to have settled into a better school match. In this scenario more experienced teachers produce better student outcomes not because experience increases productivity but because teachers settle into schools that are a better match. This suggests that those that weren't a fit are likely to have changed schools or left teaching altogether. In other words, some teacher turnover actually helps lead to a more optimal match between teachers and schools, which is good for students.
This study adds some pretty solid evidence to support giving principals more say over teacher hiring. The easier it it is to achieve these natural matches, the better it is for student performance.