Data and code for replicating analysis of Outvote's 2018 randomized study.
- outvote_data.csv: CSV containing anonymized data for all eligible subjects.
- outvote_analysis.ipynb: Jupyter Notebook that shows how to process data and replicates basic results.
The data file outvote_data.csv contains anonymized data for all n=195,118 eligible subjects. Each row contains a subject's:
- Z: assignment (1=assigned treatment, 0=assigned control)
- D: receipt (1=messaged, 0=not messaged)
- Y: voting outcome (1=voted, 0=did not vote)
- M: confident match to voter rolls (1=yes, 0=no)
- Q: position in first queue (1,2,3,... or NaN)
The refined study population consists of subjects with M=1 and Q <= 103.