Public Opinion on the Timing of U.S. School Board Elections

• In Event: School Boards, Choice, and Participation

Abstract

Educational inequality by race and social class remains unacceptably high in the U.S. (Reardon, 2011) and has grown wider after pandemic-induced schooling disruptions (West & Lake, 2021). One possible contributor is lower levels of political representation for low-income families and families of color that could help hold the school systems serving their children accountable. Indeed, Kogan et al. (2021) document the significant extent to which districts serving majority-nonwhite student populations are governed by school boards elected by a majority-white electorate.

This absence of robust representation may be due, in part, to the fact that 22 states allow city officials to schedule school board elections “off-cycle,” such that they do not coincide with national elections occurring in November of even-numbered years. Off-cycle elections tend to be low-turnout and demographically-unrepresentative events, allowing organized interest groups to dominate (Anzia, 2014; Moe, 2011). Turnout in board elections is typically below 15% (Hartney, 2021), but moving to on-cycle elections increases turnout by 25 percentage points, dwarfing the effect of reforms such as voter ID laws and mandatory early voting (Hartney, 2021). On-cycle elections also make the electorate more representative by race, age, and partisanship (Hajnal et al., 2021). While researchers have explored the impact of timing on turnout, less is known about public knowledge and opinion on election timing. We embed survey questions and experiments into nationally representative and purposefully-sampled polls of public attitudes toward education. We address the following research questions regarding school board elections:

(1) What is the public’s knowledge of when elections are held?
(2) What are the public’s preferences regarding election timing?
(3) What are the effects of messages regarding the pros and cons of on- and off-cycle elections on the public’s election timing preferences?
(4) How does opinion—and the effects of messaging—vary based on respondents’ material interests in election outcomes (e.g., whether they are district employees)?

To address these topics, we ask respondents about their knowledge and opinion of whether board elections should be held on or off cycle. We experimentally assign respondents to different versions of survey questions about their opinion on timing. Each version leads with a particular argument for or against on-cycle elections before asking the respondent’s opinion (the control condition provides no argument). We also identify subgroups with differing interests in election timing (race, education, income, parental status, age, partisanship, K-12 public school teachers, union members teachers living in the district where they are employed).

Two sources of data provide a total sample of 8,615 completed surveys. A sample of this size was key for ensuring sufficient statistical power to detect effects given multiple treatment arms (i.e., rationales). The first data source is an original survey fielded in Feb-March 2023 on the Prolific platform with two nationally representative samples (n=2,990). After maxing out the nationally representative samples we ran the survey with a larger sample of U.S. residents to increase power (n=4,000). We weight this sample to represent the nation. We surveyed an oversample of teachers (n=611) due to their unique material interest in school board race outcomes. Given it was challenging via Prolific to distinguish public K-12 school teachers from others in the education field (e.g., private schools not governed by school boards, higher education), we supplement the Prolific data with a second source—the RAND Corporation’s American Teacher Panel. We add a subset of our survey questions on election timing to their fall 2023 Omnibus Survey administered annually to a pre-verified sample of public K-12 teachers (n=1,014).

Preliminary analyses of the Prolific data suggest low overall levels of awareness regarding local election timing and majority support for on-cycle elections. Many rationales for/against on-cycle elections do shift opinion. Respondents are particularly likely to support on-cycle elections when presented with data on estimates of current voter turnout rates in school board races. Teachers seem especially susceptible to changing their opinion in favor of on-cycle elections when presented with arguments in favor of this arrangement. These effects are not moderated when teachers are presented with arguments about disadvantages of on-cycle elections for unions. Teachers’ pro-democracy sensibilities may therefore outweigh considerations of the material benefits of policy choices to them as individuals. Prior to APSA, we will have analyzed the RAND teacher data to determine whether subgroup findings replicate with a more credible sample of public school teachers. Findings have important implications for understanding why off-cycle elections persist.

Authors

• Beth Schueler, University of Virginia

• David Houston, George Mason University