With the 2024 election now appearing in our rear view window, it is time to look at what factors predicted Trump’s vote in Pennsylvania. County Health Rankings came out with their most recent (and possibly last) rankings for every county in the U.S. I am taking a look at which factors predict Trump’s % of the vote in the 2024 in PA counties. The statistics determining the ranking were compiled before the election. Two of the strongest predictors for Trump’s vote were income and the % in the county with at least some college education.
College Education
The graph above shows Trump’s percent of the vote for all 67 counties in Pennsylvania on the vertical axis. On the horizontal axis is the percent of the county with at least some college education. The graph shows that as the % with some college in the county increases, Trump’s % of the vote is likely to decrease. The R2 statistic indicates the strength of this relationship. For this graph this statistic is 55.8%. This is a fairly strong relationship. If this statistic were 100%, all 67 counties would fall on the best fit straight line.
Household Income
The second graph shows the correlation between median household income and Trump’s % of the vote in 2024. The median household income is the value that appear in the middle if they were ranked from highest to lowest. This relationship shows that as median household income increases, Trump’s % of the vote decreased in 2024.
This relationship is not as strong as the one for those with some college. The R2 statistic for this relationship is 0.333. This means that 33.3% of the variability in Trump’s vote is accounted for by median household income. This is smaller than the 55.8% in the first graph.
Philadelphia county is an outlier in both graphs. It only had 20% of the vote for Trump in 2024. In 2016, Trump had 15.37% of the vote there. It had a median household income of $60,196 and 68.8% had at least some college.
Correlations can reveal meaningful patterns that are not easy to see with the naked eye. However we should be careful about concluding that a cause and effect relationship exists among the variables. Trump’s vote being correlated with median household income and individuals with some college can always be mediated by other variables.