Panagopoulos 2006d - GOTV Street Signs

Large "Vote Tomorrow" street signs held by volunteers at high-traffic intersections are found to increase voter turnout.

Panagopoulos, Costas. 2006. "Street Fight: The Impact of a Street Sign Campaign on Voter Turnout." Unpublished Manuscript. Institution for Social and Policy Studies. Yale University.

Costas Panagopoulos reports the first randomized field experiment gauging the effects of a nonpartisan get-out-the-vote street sign campaign. As a grassroots mobilization tactic, street sign campaigns are commonly used to attract voters' attention and to stimulate participation and support. The experiment was conducted in New York City during the November 2005 municipal elections. Panagopoulos identified 14 pairs of poll sites (voting locations) that were closely matched in terms of past voter turnout. One poll site in each pair was randomly assigned to be exposed the treatment that encouraged voters to vote on the day before Election Day. The street signs treatment is found to produce higher turnout.

Electoral Context: This experiment was conducted in the context of the November 2005 New York City municipal elections. Michael Bloomberg, the incumbent Republican mayor, was seeking reelection against Freddy Ferrer, the Democratic nominee.

Subject Population: Manhattan, the city's largest borough, is comprised of 1,240 election districts that are distributed across nearly 300 voting locations (poll sites) across the city. Polling places were paired within strata using turnout in 2004 as the matching criterion. Measures were taken to avoid contamination resulting from geographical proximity of the voting locations. The author identified 14 closely matched pairs of voting locations. One poll site in each pair was randomly assigned to the treatment group and the other to the control group.

Randomization Procedure: Panagopoulos identified 14 closely matched pairs of voting locations. One poll site in each pair was randomly assigned to the treatment group and the other to the control group.

Treatment: Voters in election districts in the treatment group were exposed to volunteers holding street signs at high-traffic intersections adjacent to the voting location on November 4, 2005 between the hours of 7:00 AM and 6:00 PM. Groups of two-to-three volunteers with signs were assigned to each treatment location at all times during this period. Signs were professionally designed and produced to maximize visibility. They were 2 feet by 3 feet, with large block, dark blue lettering on a white background with the message: VOTE TOMORROW. The treatment was designed to achieve two goals: to encourage participation on Election Day and to remind voters that the election was taking place on the following day. The latter reminder was intended to reduce information costs associated with acquiring knowledge about the date of the election thereby raising the likelihood of participation further.

Findings: Regressing turnout on treatment and dummy variables for matched pairs produced an estimate for the treatment of 3.62 with a standard error of 1.70 percentage points. Additional specifications are estimated that exclude the dummy variables and include the covariate for past turnout. These results suggest a positive and statistically significant treatment effect of 3.19 percentage points (standard error = 1.72 percentage points). Panagopoulos also estimates the impact of the treatment on changes in level of turnout producing results which suggests a positive treatment effect of 2.7 percentage points with a standard error of 1.84 percentage points. The conservative OLS estimate suggests that the sign treatment produced each additional vote for approximately $19.

Tiffany Davenport – Fri, 2006 – 09 – 22 11:26