Angrist et al 2010 — Inputs and Impacts in Charter Schools KIPP Lynn

Summary

Angrist, Dynarski, Kane, Pathak, and Walters (2010) evaluate KIPP Academy Lynn — a "No Excuses" charter middle school in Lynn, Massachusetts serving a predominantly Hispanic, low-income population — using an applicant lottery as an instrument for attendance. The school is oversubscribed, and lottery winners and losers are otherwise observationally equivalent, allowing lottery offer to serve as a valid instrumental variable (IV) for KIPP enrollment. The paper finds large causal gains in math (0.35σ per year) and more modest but significant reading gains (0.12σ per year), with reading improvements concentrated almost entirely among Limited English Proficiency (LEP) and special-education (SPED) students. The results contradict the critique that high-performing charter schools achieve their results by cream-skimming; the KIPP Lynn complier population starts below the state average in 4th grade and includes high proportions of LEP and SPED students.

Key Claims

• KIPP Lynn applicants enter 4th grade roughly 0.3 standard deviations below the Massachusetts state mean in math, confirming the school serves a genuinely disadvantaged population, not a selected high-ability group.
• The causal effect of attending KIPP Lynn is +0.35σ per year in math and +0.12σ per year in reading, identified by lottery offer IV; these are large by the standards of education interventions.
• Reading gains are heterogeneous: the overall 0.12σ average masks ~0.30–0.35σ gains for LEP and special-education students and near-zero gains for other students. This heterogeneity pattern suggests KIPP's extended-time, intensive-literacy model is especially effective for students with academic or linguistic barriers.
• Math gains are broad-based across the student distribution, not concentrated in particular subgroups.
• The No Excuses school model — 10.5-hour days, August start, Saturday school, strict behavioral code, binding parent/student commitment contracts — is the set of inputs examined; the paper establishes that this bundle produces achievement gains but does not attempt to decompose which specific input drives effects.

Concepts Introduced or Extended

• Charter School Effectiveness
• Heterogeneous Treatment Effects
• Local Average Treatment Effect

Entities Mentioned

• Joshua Angrist
• Susan Dynarski
• Thomas Kane
• Parag Pathak
• Christopher Walters
• KIPP

Quotes

"KIPP Lynn applicants are disproportionately Hispanic and include large shares of limited English proficiency (LEP) and special education students. These groups are often underserved by urban public schools and, critics charge, by some charter schools."

"The results suggest that the time-intensive No Excuses model can produce large achievement gains, at least in middle school math, for urban students who start well below grade level."

My Take

A clean and influential lottery-IV study. The short P&P format (5 pages) limits the depth of input analysis — the paper establishes the causal effect of the KIPP bundle but cannot separate extended time from discipline model or teacher quality. The subgroup finding (reading concentrated in LEP/SPED) is the most substantively interesting result and arguably more important than the headline average, since it directly addresses the sorting critique. The companion paper (JPAM 2012, "Who Benefits from KIPP?") extends the analysis with more detail. A limitation: KIPP Lynn is one school; generalizability to the broader KIPP network or No Excuses charters in other cities requires the subsequent replication literature (Dobbie and Fryer 2013; CREDO studies).