Straighten Up and Fly Right: The Impacts of Charter Schools on Behavioral and Civic Outcomes (North Carolina charter schools and voting)

North Carolina administrative data linked to voter files; exact n not extracted

North Carolina charter and traditional public school students.

Matching and regression on administrative data rather than a lottery. Cohodes and Feigenbaum's critique is the decisive one and is worth recording verbatim in spirit: matching on pre-treatment test scores may work for test-score outcomes, but for voting there is no appropriate pre-treatment variable, so unobserved antecedents of civic behaviour are not absorbed. TITLE CAVEAT: the exact registered title was not verified, so no DOI is attached and the recorded title should be treated as approximate until checked.

North Carolina charter attendance is associated with higher voter registration and turnout, plus better attendance, fewer suspensions and less criminal activity. Consistent in sign with the lottery evidence, but not identified.

Not assessed in detail; see exclusion note.

Its direction agrees with the two charter LOTTERY studies (Gill et al. 2020; Cohodes and Feigenbaum), but Cohodes and Feigenbaum explicitly note that matching cannot be trusted for non-academic outcomes because there is no pre-treatment analogue of voting to match on.