Ballot order is close to random, and Shi and Singleton turned that quirk into a natural experiment. When a candidate's name lands higher on the ballot, that candidate wins more often for reasons unrelated to platform, competence, or voter intent. The authors use that lift to isolate what happens when an educator - a current or former teacher - joins a school board rather than a non-educator. The study earns a PEARLS review because it is one of the few pieces of work that estimates a causal effect of board composition rather than a correlation, and because its headline result cuts against a common assumption: that putting teachers on boards improves classroom outcomes. For a full explanation of the PEARLS framework used in this review, see our [methodology page](/methodology.html).
The available excerpt does not report an exact count of districts, elections, or board seats, which is a real gap for a reader trying to gauge precision. What can be said from the design: the unit of analysis is board members elected across districts where ballot order was randomized, and the sample is restricted to jurisdictions that both randomize ballot position and hold the kind of contested multi-candidate races the method requires. That restriction matters. States and counties that randomize ballot order are not a random draw of American districts, so the sample skews toward the electoral systems that permit this identification strategy. A practitioner in a district with fixed or incumbent-first ballot order is not looking at a mirror of their own governance environment. Judge the representativeness as adequate for the mechanism being tested and limited for national generalization.
The design is quasi-experimental, specifically an instrumental-variables strategy built on randomized ballot order. This is the right tool for the question and a meaningful step above the observational studies that dominate this literature. Because ballot position is assigned at random and plausibly affects only vote share and not a candidate's underlying qualities, it functions as an instrument for whether an educator wins a seat. That isolates the effect of educator presence from the confounders that wreck ordinary regressions - districts that elect teachers differ from districts that do not in dozens of unobserved ways, and a naive comparison cannot separate the board member from the community that chose them.
The execution depends on two conditions worth naming. First, the ballot-order effect must be strong enough to move seat outcomes, or the instrument is weak and estimates become unstable. Second, ballot order must not influence district outcomes through any channel other than who wins, an exclusion restriction that is credible here but not directly testable. Published in a peer-reviewed economics journal, the work cleared serious methodological scrutiny.
The authors assert causation, and the design earns it for the outcomes where estimates are precise. Adding one educator to a board reduces charter schooling in the district and raises teacher salaries relative to what other board members deliver.
On student test scores, the estimates are negative but imprecise - meaning the confidence intervals are wide enough that the true effect could be a small decline, zero, or a modest gain. The responsible reading is that the study does not detect a clear achievement effect. The authors report the imprecision plainly rather than spinning a null into a finding.
The study was published in 2023. Charter authorization fights and teacher-pay disputes remain live in most states, so the outcomes measured are the outcomes boards are still voting on. The relevance qualifier is structural, not temporal. The findings travel cleanly to districts that resemble the sample - elected boards with authority over charter authorization and salary schedules. They travel poorly to appointed boards, to states where charter authorization sits with a separate body, or to districts where salary schedules are set through statewide bargaining rather than local negotiation. A practitioner should first ask whether their board controls the levers this study measures.
The excerpt does not specify a funding source. Publication in the American Economic Journal: Economic Policy carries a disclosure requirement, so a funded relationship would ordinarily appear in the article's acknowledgments, and readers should check that section directly rather than take its absence here as proof of independence. On its face this reads as academic work with no obvious sponsor stake in the direction of the result.
The study is deliberately narrow, and that is a strength for causal precision and a limit for breadth. It measures a single dimension of board composition - educator versus non-educator - and a short list of outcomes: charter presence, teacher salaries, and test scores. It does not address board racial or ethnic composition, governance behavior, meeting conduct, or superintendent relations.
Placed against neighboring work, this is not a lone finding.
Board members can conclude that who sits on the board changes fiscal and structural decisions - electing an educator tends to lift teacher pay and reduce charter presence where the board controls those levers. They should not conclude that adding educators improves or harms student achievement, the study does not detect a clear effect either way.
Three concrete moves. Confirm your board controls charter authorization and salary schedules before applying these estimates. Treat occupational composition and racial-ethnic composition as separate questions. When someone cites this paper to argue educators hurt test scores, ask them to quote the confidence interval.
Weight this study heavily for its fiscal and structural claims. Weight it lightly on achievement, where the design produced a genuine non-answer.
Found a study on school board governance? Paste the DOI below and we'll add it to the PEARLS review queue.
We'll evaluate it using the PEARLS framework and post the review when complete.