13 - A Randomized Control Trial of Pediatrician-Led Voter Engagement

Publication Number: 4012.13

Margaret N. Jones, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States; Kelly Hunter, Northwestern University, Chicago, IL, United States; Sam Fuller, Harvard University, Cambridge, MA, United States; Daniel B. Magleby, Binghamton University, SUNY, Binghamton, NY, United States; Andrew F. Beck, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States; Tierra M. Dennis, CCHMC, Cincinnati, OH, United States; Hilary Izatt, Binghamton University (SUNY), Binghamton, NY, United States

Background: There is increasing evidence that voting is tied to improved health outcomes, yet voter registration and turnout is low in the United States. Healthcare-based voter engagement has been proposed as a promising strategy for promoting safe and equitable access to voting. To date, no randomized controlled trial (RCT) has tested the impact of healthcare-based voter engagement on voting behavior. Studies have shown that outreach from known sources improves civic engagement.

Objective: To estimate the causal effect of brief pediatrician-led conversations about voting on household voter turnout.

Design/Methods: We conducted an RCT at a single academic primary care office at our institution. Patients seen October-November 2024 were randomized using in-clinic die-roll assignment to either the control group (exposed to passive voting signage in exam rooms, Figure) or the intervention group (exposed to signage, and received a brief, nonpartisan recommendation from their pediatrician about the importance of voting). Outcomes included household turnout (whether an adult at the patient's address cast a ballot) in the 2024 U.S. general election via address linkage to state voter files.

Results: There were 2,078 households included in the analysis, with 146 households in the treatment group and 1,932 in the control group. Demographics for individuals from the included households are shown in Table 1. There were no significant demographic differences between the treatment and control groups. Overall, 46% of included households had at least one adult vote in the 2024 general election, while the state-wide voter turnout for eligible voters in our state was 65%. There was no significant difference in voter turnout between the treatment and control groups (estimated effect = 1.6%, p = 0.71, Table 2).

Conclusion(s): We successfully conducted a rigorous RCT of pediatrician-led voter engagement integrated into a clinical setting. The voter turnout rate for this academic primary care pediatric office patient population was notably lower than the overall voter turnout rate in our state. In this small-n study, there was no significant difference in voter turnout between households who received a brief verbal recommendation from their pediatrician and those who were exposed only to passive voting signage. Future analyses should investigate the effectiveness of alternative approaches (e.g., flyers/postcards, badge cards, recommendations from interdisciplinary healthcare staff, etc.) for healthcare-based voter engagement.

Figure. Voting signage hanging in every primary care exam room
Institution name obscured for review

Table 1. Characteristics of individuals from included households


Table 2. Ordinary least squares estimates from a bivariate regression of household turnout on treatment status. The analytic sample includes n = 2,078 households (treatment = 146, control = 1,932). The dependent variable equals 1 if at least one household member voted in the 2024 U.S. general election.
Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001.
+Interpretation: Treated households were 1.6% more likely to vote than households in the control group (p = 0.707)