Mining Voter Behaviour and Confidence: A Rule-Based Analysis of the 2022 U.S. Elections

This study explores the relationship between voter trust and their experiences during elections by applying a rule-based data mining technique to the 2022 Survey of the Performance of American Elections (SPAE). Using the Apriori algorithm and setting parameters to capture meaningful associations (support $\geq \mathbf{3 \%}$, confidence $\geq 60 \%$, and lift $>1.5$), the analysis revealed a strong connection between demographic attributes and various voting-related challenges, such as registration hurdles, accessibility issues, and queue times. For instance, respondents who indicated that accessing polling stations was “very easy” and who reported moderate confidence were found to be over six times more likely (lift $=6.12$) to trust their county's election outcome and experience no registration issues. A further analysis, which adjusted the support threshold to 2 %, specifically examined patterns among minority voters. It revealed that 98.16% of Black voters who reported easy access to polling locations also had smooth registration experiences. Additionally, those who had high confidence in the vote counting process were almost two times as likely to identify as Democratic Party. These findings point to the important role that enhancing voting access and offering targeted support can play in building trust in the electoral system, particularly among marginalized communities.