Cost of Voting and State Politics

Examining the relationship between state voting laws and political sentiment.

The Cost of Voting Index is an ongoing analysis by Michael J. Pomante II, Ph.D., a political scientist. It “examines election laws and policies and calculates a single measure of the relative difficulty of voting for each state. States with smaller values make voting more accessible than states with larger values.” Data and accompanying peer-reviewed papers are available on the Cost of Voting Index website.

Briefly, “The COVI measures the time and effort required to vote, encompassing criteria such as registration requirements, voter identification laws, early voting availability, and absentee voting procedures.” ⁽¹⁾Pomante II, Michael J. (2024). “The Cost of Voting in the American States: 2024.” Election Law Journal: Rules, Politics, and Policy. DOI: 10.1089/elj.2024.0037. Available from https://costofvotingindex.com/publications I’ve taken the data for 2024 and plotted it against the percent of Donald Trump voting in the states. Chart 1 shows this. (click for larger charts)

I asked ChatGPT to perform a regression analysis and explain. The summary was: “This analysis examines how the difficulty of voting (measured by the “Cost of Voting Index”) varies with a state’s political leanings and population. The results suggest that states with a higher percentage of Trump voters in 2024 tend to have stricter voting laws (higher cost of voting). However, state population size does not seem to have a meaningful impact on voting difficulty.” In more detail:

We performed a multiple linear regression to examine the relationship between:

• Independent variables:
  • Share of Trump Vote in 2024 (X1)
  • State Population (X2)
• Dependent variable:
  • Cost of Voting Index (Y)

Key Findings from the Regression:
– R-squared (0.425): This indicates that approximately 42.5% of the variation in the Cost of Voting Index can be explained by the Share of Trump Vote and State Population.
– Share of Trump Vote (Coefficient = 8.5481, p-value < 0.001): A statistically significant predictor of the Cost of Voting Index, suggesting that as the share of Trump votes increases, the Cost of Voting Index also tends to increase.
– State Population (Coefficient = 0.8143, p-value = 0.751): This variable is not statistically significant, meaning there is no strong evidence that population size affects the Cost of Voting Index.

I also examined voter turnout rate as a function of the Cost of Voting Index. Chart 2 illustrates this:

As before, I asked ChatGPT for a regression analysis: “Think of this as trying to understand what factors influence how many people show up to vote in each state. We looked at whether making voting harder (higher Cost of Voting Index) or having a larger population affects turnout. The results suggest that voter turnout may slightly drop when voting is harder, but the effect is weak. However, the population of a state does not seem to play a major role in voter turnout. The model itself doesn’t explain much of the variation, meaning there are likely other factors (like political engagement, campaign efforts, or demographics) that influence voter turnout more strongly.” In more detail:

We performed a multiple linear regression analysis to determine the relationship between Voter Turnout (dependent variable) and two independent variables:

1. Cost of Voting Index (measuring how easy or difficult it is to vote in a state)
2. State Population (the number of residents in the state)

Key Findings:
– The R-squared value is 0.068, meaning that only 6.8% of the variation in voter turnout is explained by the Cost of Voting Index and State Population.
– The Cost of Voting Index has a coefficient of -0.0123, suggesting that as voting becomes harder (higher cost index), voter turnout decreases slightly.
– The State Population has a coefficient of -6.35e-10, indicating that population size has a negligible impact on voter turnout.
– The P-value for Cost of Voting Index is 0.077, which means it’s not statistically significant at the typical 0.05 threshold but is close. This suggests a possible weak relationship between the difficulty of voting and turnout.
– The P-value for State Population is 0.591, meaning there is no significant relationship between state population and voter turnout.