Block 4: Randomized trial: social pressure
Week 3 · Causality I · Imai §2.4.2 · about 4 min
NoteHow this works
Run each cell with Run Code, look at the output, then answer the question at the end of the block. Every cell is editable; Start Over resets it. Errors are information: read them.
Gerber, Green & Larimer: before the 2006 primary, about 300,000 registered voters were randomly assigned to different mailings. Outcome: did they vote (primary2006)? Control = no mailing; Civic Duty = “do your civic duty”; Hawthorne = “you are being studied”; Neighbors = shows the household’s and neighbours’ voting records.
The data are big (300,000 rows): loading takes a few seconds.
1. Load and inspect
How many households per group? Why is Control so much larger?
2. Turnout in each group
Conclusion: Compared with the control group, which mailing raises turnout most, and by roughly how much?
Turnout is about 29.7% in the control group; Neighbors raises it to 37.8%, about +8.1 points. Civic Duty (+1.8) and Hawthorne (+2.6) work much less.