Due Friday, March 29, at 5pm Central time.
As before, please turn in your answers in a single notebook, emailed to Diana Dewalle.
Problem 1 (55 Points): Legislative Discrimination
Go onto google scholar and find Daniel M. Butler & David E. Broockman, "Do Politicians Racially Discriminate Against Constituents? A Field Experiment on State Legislators" 55 American Journal of Political Science 463 (2011).
Problem 1.A (20 Points):
Explain their results in plain english. What were their methods? What did they find?
Problem 1.B (15 Points):
Contrast their methods to the methods used by the Metropolitan Milwaukee Fair Housing Council (MMFHC) in United States v. Balistrieri, from our week 7 readings, to check for discrimination. Identify at least one advantage and one disadvantage (in terms of reliably identifying discrimination if it exists and not identifying discrimination if it doesn't exist) that the Butler/Broockman method has relative to the MMFHC method.
Problem 1.C (20 Points):
Attempt to replicate some of their core results. Their dataset is on our class server as Butler_Broockman_AJPS_2011_public_csv.csv
Notes:

You won't be able to replicate any of the results from the appendix, because privacy rules kept them from sharing the data from the control variables.

I know that the paper isn't terribly clear on what statistical tests they used to generate the p values in the tables. Come up with your own!

You don't by any means need to replicate all or even most of their results. I'll be satisfied if you plausibly use the data provided to convince a reader that the main findings of the paper (as you defined them in subproblem A) are correct.
Problem 2 (45 Points): Faculty Discrimination
Download the dataset prof_sex_discrim.csv
from our class server. It's a dataset of "faculty in a small midwestern college" allegedly from a sex discrimination case. (The dataset is swiped from a classic text, Weisberg, Applied Linear Regression, and from a professor at Princeton, Germán Rodríguez, who made the data available online. The columns should be fairly selfdescriptiveit's all faculty who are tenured or on the tenure track, and there are variables for salary, rank, years in rank, years since degree, and whether they have a doctorate or just a masters.
I want you to explore this dataset in a fairly openended fashion to see whether you can make any conclusions about the presence of sex discrimination. Here are the rules:

Use at least one data visualization, and you have to explain, in words, what you see. The more informative, the more points.

Use at least one statistical test, and you have to explain, in words, why that statistical test makes sense as a way to tell whether this school is discriminating. Again, the more informative, the more points.

Identify at least one potential problem with the statistical analysis you ran, and explain why, given the data you have (as identified by visualization or exploratory data analysis), this problem should worry you. Suggest at least one technique you might use to alleviate or avoid this problem.
Each of those rules is worth 1/3 of the points for this problem (15 points).