Main Regression

The most important thing to note about this table is the relative strengths of the coefficients. The number of 25-34 year olds, which probably received the most amount of attention from the marketing literature, proves to be insignificant at even a 5% level in two of the three models in which it was featured. The one where it is significant is when the white population coefficient is swapped out for the total population coefficient, at which point it becomes significant at a 5% level. This seems to suggest that the young people in highly populated (urban) areas are somewhat relevant, regardless of the racial makeup of the areas they inhabit. The use of median age as an alternative coefficient does not yield any more meaningful results.

The white population seems to do even more poorly, being statistically insignificant in all three of the results where it is run. Its alternative, total population, is significant at a 5% level, but actually shows a negative coefficient once income, age, and liberalism have been controlled for. While not particularly immense or significant, there is a change in the sign of the white population coefficient when income is not controlled for.

Overall, the two coefficients that seem to be strong in the models are median income and number of Obama voters. In all the models that feature both, median income tends to be slightly stronger as a coefficient, while Obama voters are consistently the most statistically significant coefficient.

Table 1.1: Main Regression Results