Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We investigate a quasi-randomized field experiment to determine the effectiveness of door-to-door campaigning by a candidate on outcomes in a general election. The candidate campaigned door-to-door in quasi-randomly-selected voting precincts, which represent the treatment group. Precincts in which the candidate did not campaign door-to-door are the control group. We utilize detailed precinct-level demographic characteristics, vote totals for the candidate in the prior election, and vote totals for a candidate for another office that appeared on the same ballots in both elections. This information allows us to control for various factors using difference-in-difference and first difference regression models to account for differences between treatment and control groups. We find that door-to-door campaigning by the candidate increases the candidate's vote by 3 percentage points and the vote margin by 6 percentage points in a two-candidate race.