library(ggplot2)

library(ggthemes)

dat <- read.csv("/Users/jefflauer/Desktop/hate_crimes.csv")

ggplot(data = dat, aes(x = median_household_income, y = share_voters_voted_trump, color = share_white_poverty)) +
    geom_point(size = 2, alpha = .8) +
    geom_text(label = dat$state, size = 2, hjust = 0.5, vjust = -1, alpha = .5) +
    labs(title = "Comparing Income to Voting for Trump", x = "Median Household Income by State", y ="Share of Voters Who Voted for Trump", colour = "Share of White Poverty") +
    coord_cartesian(ylim = c(.25, 0.8)) +
    theme_gdocs() 

ggplot(data = dat, aes(x = median_household_income, y = share_voters_voted_trump, color = share_white_poverty)) +
    geom_point(size = 2, alpha = .8) +
    geom_text(label = dat$state, size = 2, hjust = 0.5, vjust = -1, alpha = .5) +
    labs(title = "Comparing Income to Voting for Trump", x = "Median Household Income by State", y ="Share of Voters Who Voted for Trump", colour = "Share of White Poverty") +
    coord_cartesian(ylim = c(.25, 0.8)) +
    theme_gdocs() + 
    geom_smooth(method = "lm")

I found this data set through FiveThirtyEight. While it included a lot of different variables, including statistics on reported hate crimes and unemployment, I focused in on Median Household Income, represented by the x axis, the share of votes that went to Trump in 2016 by each state, represented by the y axis, and the concentration of white poverty by each state, represented by the shade of blue. Light blue indicates high concentrations of white poverty, and dark blue indicates low concentrations of white poverty. As you can tell by the second graph, there seems to be somewhat of a correlation between states with a low median household income and also having a larger percentage of voters who went for Trump. This can be seen in states like West Virginia, Mississippi, and Alabama. Conversely, states like Hawaii, Maryland, and Massachussetts with high median household incomes had very low percentages of Trump voters. Just by eyeballing these two graphs, it appears that both of these variable correlate with concentrations of white poverty, which is significant given that Trump, like most Republicans, was elected by a largely white constituency.

I think there are a number of different points you could pull from these graphs. On the one hand, I think some would argue that given Trump’s xenophobic rhetoric, poor white voters in impoverished states are galvenized by racist talking points and policy proposals. However, I think more in depth data on the specific voting behavior as well as perhaps more qualititaitve research about impoverished white people in these states would have to first be conducted before we can attest to the validity of that point. Nevertheless, there does seem to be a significant economic draw to voting for Trump. Given that states with low median household incomes went strongly for Trump, while states with higher household incomes did not, this suggests that Trump’s campaign and talking points may have addressed economic concerns that were of interest to poor and working class white voters. Whether or not Trump’s actual policies benefit this voting demographic is an entirely different claim and would require further research. Regardless, support for Trump does seem t be correlated to states with lower household incomes, specifically amongst impoverished white populations.