Are KIPP schools the solution to what ails the poorest, most disadvantaged urban students? The KIPP educational paradigm rests on a few core principles (“the Five Pillars”) – High Expectations, Choice and Commitment, More Time, Power to Lead, and Focus on Results. Students at KIPP schools work in highly structured learning environments for much longer hours and are generally led by young, talented teachers. At the administrative level, principals are given a freer hand to achieve results without bureaucratic or teacher unions constraining them. But to keep their affiliation, KIPP headquarters requires that schools prove that they are achieving measurable results. Long-term funding comes primarily from the local school districts so per student costs are comparable to public schools.
Although KIPP schools have their fair share of admirers, including education secretary Arne Duncan, they also have their skeptics. One criticism is that KIPP schools “cream skim”, admitting and retaining the most talented students in disadvantaged communities. A slightly different criticism is that even if they don’t cream skim, KIPP schools only benefits motivated students and parents, and that the less high achieving enrollees fail to benefit. These assertions seem plausible, but they have not been empirically tested. In a new paper, Josh Angrist and colleagues in JPAM debunk both claims in the case of one KIPP school in Lynn, Massachusetts.
The main contribution of the paper is its strong quasi-experimental design: the authors take advantage of the fact that Massachusetts laws requires charter schools that are oversubscribed to admit students using a lottery. This allows them to compare standardized test results for lottery winners and losers, where the only distinction between these groups should be whether they happened to randomly win the entrance lottery. This “intent to treat” analysis also improves upon other studies that retrospectively compare the test results ofKIPPenrollees and public school students, since retrospective studies cannot fully control for differential enrollment and drop-out patterns.
The results indicate that KIPP substantially improves student performance in math and language (average achievement gains of 0.36 standard deviations in math and 0.12 standard deviations in reading for each year spent at KIPP). The greatest gains came for students with identified learning disabilities or language barriers. The effect by subgroup estimates are shown below (in small format), and they show for example, that after adjusting for demographics and baseline scores, that limited English proficiency students experience 0.47 standard deviations in math improvement.
We may dismiss the results of this study as a statistical fluke from a single KIPP school – which may not represent the broader experience of these kinds of schools nationally – but the results are actually consistent with a growing literature on results-oriented charter schools. For example, Roland Fryer and Will Dobbie show that the Harlem Children Zone produces outstanding achievement gains (essentially bringing some of the lowest performing black students in NYC up to the average level of white students).
We still do not have a good grasp on is what makes these schools successful. One answer may be everything: more hours, stricter standards, better teachers, and more parental involvement could all help to raise achievement, but some may yield a much greater bang for the buck. In particular, if the answer is that we need KIPP quality teachers in all urban schools, then it may take years to build up the pipeline to train and retain excellent teachers. In the meantime, there may be some simpler tools that can diffuse more widely to disadvantaged students like using small financial rewards or creating school-wide positive behavior programs.