While polling and forecasts predict the winner of the Presidential Election based on past voting preferences and early ballot preferences, one professor is using a different metric – Music choice.

Jeff Smith, a professor of data science and music at Stanford, investigated whether political preference would correlate with musical taste. To do so, they looked at geographical data, voting history from previous presidential races, and genre tastes of individuals, to find patterns.

Through this analysis, they found strong correlations between the three measures – for example, perhaps more obviously, people who listened to Christian and country music were more likely to live in historically Republican-voting states, while people listening to Latin or Broadway music were more likely to live in Democrat-leaning states.

Using these preferences as a basis, Smith and his team investigated how swing states were likely to vote, which differed to the majority of current polls. As an example, Smith argues that Nevada and Florida will lean Republican whilst early turnout data suggest they will vote Democrat. PPE and HSPS applicants should consider how valid early turnout data has been in predicting election outcomes, and to what extent forecasting is a useful measure of presidential outcomes, particularly for swing states.

More Resources