Collective Bargaining of Charter School Principals and School Performance

ABSTRACT

Because charter school principals are granted more autonomy and discretion than principals of traditional public schools, it is imperative to search for the attributes of principals that may improve charter school performance. This study examines the relationship between principals’ collective bargaining and charter school effectiveness. Using propensity score matching based on the School and Staffing Survey, I find that principals’ unionization is positively and significantly associated with charter school performance. I also identify potential mechanisms through which principals’ unionization influences charter schools: higher pay and better working conditions for principals, more formal principal evaluation, and more informal teacher evaluation.

Acknowledgments

We thank the National Bureau of Economic Research (NBER) for providing us with the necessary facilities and assistance. We also thank the National Center for Education Statistics (NCES) for kindly providing us with the data. The views expressed herein are our own and do not necessarily reflect the views of the NBER or the NCES.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. It is noteworthy that the quality and the stakes associated with the large-scale assessment for school performance have been established only after the No Child Left Behind Act of 2001 (NCLB). Thus, studies conducted before and after NCLB may not have the same qualitative standard and approach.

2. To compensate for over-sampling and under-sampling of the stratified survey design, and to obtain unbiased estimates of the national population of schools and principals, each observation is weighted by the inverse of its probability of selection during each survey year. To compute final sampling weights, a bootstrap variance estimator was used. The bootstrap variance reflects the increase in precision due to large sampling rates because the bootstrap is done systematically without replacement, as was the original sampling. The bootstrap replicates basic weights (inverse of the probability of selection during each survey year) were subsequently reweighted.

3. Adequate Yearly Progress (AYP) is a measurement of annual achievement for public schools and districts mandated by the U.S. Department of Education under the establishment of the No Child Left Behind Act (NCLB) of 2001. The goal of AYP is to determine how every public school and school district performs academically, according to standardized test results, and to assure that all students meet or exceed state standards within a 12-year timeframe, starting with the 2002–03 academic school year. To meet AYP, districts, schools, and student groups were expected to meet three sets of requirements: (i) achieve 95% student participation rate on statewide tests, (ii) demonstrate growth in percentage of students scoring at the proficient or above level in English language Arts (ELA) and Mathematics on statewide tests, and (iii) meet established graduation rate targets, if applicable.

4. Alternative schools offer a curriculum designed to provide alternative or non-traditional education; these schools do not specifically fall into the categories of regular, special program emphasis, special education, or vocational school.

5. Primary schools are schools with at least one grade lower than 5 and no grade higher than 8. Middle schools are schools with no grade lower than 5 and no grade higher than 8. High schools are schools with no grade lower than 7 and at least one grade higher than 8. Combined schools are schools with at least one grade lower than 7 and at least one grade higher than 8.

6. Both logit and probit models produce similar estimates with the linear probability model, so I report results from the linear probability model for easier interpretation of coefficients.

7. The percent of bias is the percent difference of the sample means between the treated group and the non-treated group, and it is calculated as a percentage of the square root of the average of the sample variances in the treated and non-treated units, which is the formula from Rosenbaum and Rubin (1985).