Student-based door-to-door canvassing during a highly competitive election season does not increase turnout in the overall subject population. However, student canvassers are successful in increasing turnout among voters under the age of thirty.

Bennion, Elizabeth A. 2005. "Caught in the Ground Wars: Mobilizing Voters During a Competitive Congressional Campaign. The Science of Voter Mobilization. Special Editors Donald P. Green and Alan S. Gerber. The Annals of the American Academy of Political and Social Science. vol. 601: 123-141.

Elizabeth Bennion examines the effectiveness of a student-based, nonpartisan voter mobilization coalition conveyed nonpartisan get-out-the-vote messages through door-to-door canvassing in three South Bend precincts on the weekend before the November 2002 election. These get-out the vote efforts took place during a competitive election season-one that included door-to-door partisan campaign efforts by interest groups, political parties, and candidate campaigns. The experiment was designed to test the effect a nonpartisan mobilization campaign can have in the midst of door-to-door campaigning by both major political parties. As the initiative was carried out by college students, the group was particularly interested in the effects of their efforts on younger voters. While the nonpartisan mobilization campaign did little to increase the likelihood of voting among older voters, it had a strong effect on voters younger than thirty-the voters least likely to be contacted by partisan campaigns.

Electoral Context: This experiment was conducted during the weekend before the November 2002 election for Indiana's Second Congressional District in South Bend, Indiana. The election was very competitive and characterized by many overlapping door-to-door partisan canvassing, phone drive and mail mobilization initiatives by interest groups, political parties and candidate campaigns. Statewide, 38 percent of voting age citizens voted in the November 2002 election. In the St. Joseph's County where this experiment was conducted, 46 percent of the voting age population voted.

Subject Population: Canvassers targeted registered voters in three precincts in the predominately white, working class South Bend neighborhood of River Park.

Randomization Procedure: 2,178 individual registered voters were randomly assigned to the treatment or the control group. Individuals in the treatment group were then randomly assigned to receive one of two flyers along with the face-to-face presentation of an accompanying script emphasizing different reasons to vote.

Treatment: Individuals in the control group were not contacted. Half of the members of the treatment group were randomly assigned to receive literature and hear an accompanying canvassing script emphasizing a civic duty to vote and the other half of the treatment group received a flyer and message emphasizing the closeness of the election. Canvassers introduced themselves, reminded subjects about the election and concluded with either the civic duty or the close election message. They then handed the subject the flyer that corresponded with the script.

Findings: Among the entire sample, the canvassing effort produced no statistically significant effect on turnout. Without controlling for past voter turnout, the treatment-on-the-treated effect of the GOTV effort was .6 percentage points (standard error: 5.1 percentage points). Controlling for voter history, the canvassing effort increased the probability of turnout by 1.8 percentage points with a standard error of 4.1 percentage points. Separating the sample in to age groups under and over thirty, it is evident that the canvassing effort had no effect on voters over the age of thirty. The estimate of the treatment effect on treated voters over thirty years old is negative 2.0 percentage points with a standard error of 4.5 percentage points. However, of note is the fact that the treatment effect on treated voters between the ages of eighteen and twenty nine was 18.1 percentage points with a standard error of 9.9 percentage points. For the entire sample, there was a small, statistically insignificant difference in the effects of the civic duty and close election messages. Voters under the age of thirty, however, were particularly motivated by the civic duty message. Contact with the civic duty message among young voters was 8.7 percentage points (standard error = 6.5 percentage points) more effective in mobilizing turnout than being contacted with the close election message.