# S670 Problem set 2

\$35.00

## Description

The data
The file anes pilot 2016.csv contains data for 1200 respondents to the 2016 American National
Election Studies (ANES) pilot study conducted in January 2016. To simplify things, we assume
the respondents are a simple random sample from the population of adult U.S. citizens (this is not
true but is close enough for our purposes.)
The data includes “feeling thermometer” measurements for various public figures. Respondents
give a score from 0 to 100 for the figure, where 0 indicates “very cold” and 100 indicates “very
warm.” (A value of more than 100 denotes a missing value and should be omitted.) We’ll look at
the feeling thermometer scores for four Presidential candidates: Donald Trump (variable fttrump),
Hillary Clinton (fthrc), Bernie Sanders (ftsanders), and Marco Rubio (ftrubio).
In addition, we know (in retrospect) that immigration was a decisive issue in the 2016 election.
One of the questions asked in the pilot study was “Should the number of people who are allowed
to legally move to the United States to live and work be increased, decrease, or kept the same as
it is now?” The responses, recorded in the variable immig numb, were coded numerically:
• 1: Increased a lot
• 2: Increased a moderate amount
• 3: Increased a little
• 4: Kept the same
• 5: Decreased a little
• 6: Decreased a moderate amount
• 7: Decreased a lot
For the purpose of this problem set, treat this as an (ordered) categorical variable.
The questions are on the next page.
1
Questions
1. Draw ONE graph that clearly shows the differences in the shape of the distributions of feeling
thermometer scores for Clinton, Sanders, Rubio, and Trump. (A faceted plot counts as one
graph.) Describe what you see in a paragraph.
2. Draw ONE graph that clearly shows the differences in the mean feeling thermometer score
for each level of immigration attitude for each of the four candidates. (Again, a faceted plot
counts as one graph.) Describe what you see in a paragraph.
You’ll be graded on both graphs and writing. All graphs should be sufficiently labeled such
that if they’re taken out of context, a viewer should have a chance of working out what they are