Problem Set 9

How does income inequality affect voting patterns?

We will do a brief investigation on the association between income inequality and vote share for a specific party. Does greater income inequality lead to voters favoring a certain party?

Datasets

• National inequality dataset used in Lab 9 (national_ineq_union)
• National Presidential Election Data (elections_national)
• The metro-level dataset constructed in Lab 9 (msa_inq_union)
• A metro-level dataset on voting in the 2016 presidental election (election_msa)

The metro-level election data began with county-level data and require a cross-walk to convert it to metro-level data. There are - multiple counties to a metro area, so we aggregated counties to the metro level. If you are interested in doing this yourself, you’ll need a metro-to-county cross walk or CBSA to county crosswalk

Questions

1. National Level trends on income inequality

Link the national level datasets together. Plot a figure that shows how income inequality and vote share for either the Democrats or Republicans change between the years 2000-2016. (Hint: You will have to create a new variable for vote share).

-Hint 1: There is only voting data for national election years (2000, 2004, 2008, etc . .). This creates a problem with you try to plot the data. See how others deal with this using na.omit on stackoverflow

-Hint 2: For the voting data, it is useful to combine geom_line with geom_point. Why? `geom_point’ will show on the figure which years you have data for.

Are there patterns over this time period between inequality and vote share?

2. Map of Vote Share

Link the metro-level datasets. Create a map that visualizes vote share for metro areas. Can you describe which metro areas learn more democratic or republican?

3. Relationship between inequality and vote share

Using the metro-level data, generate a scatterplot with income inequality on the x-axis and vote share (for eithe party) on the y-axis. Add a linear regression line.

Is there any pattern between income inequality and vote share for either major party?

4. Linear Regression

Estimate the association between income inequality and vote share using linear regression. The outcome varible should be vote share (of either party) and the indepdent variable is top 1% share. Do a series of regressions, where you begin to control for metro-level variables.

How does the intial regression between vote share and income equality look like? What happens when you add controls? Based on these results, what can you infer?