Trivedi 2005 - Ethnic Identity-Based Direct Mail
Three variations of identity-based direct mail are found to have little effect on turnout among Indian-Americans.
Trivedi, Neema. 2005. "The Effect of Identity-Based GOTV Direct Mail Appeals on the Turnout of Indian Americans." 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: 115-122.
Neema Trivedi presents the results of a randomized field experiment testing the effectiveness of three different identity-based appeals communicated via direct mail to registered Indian American voters in Queens County, NY.
The appeals made salient different identities: the voter as a U.S. citizen, a U.S. citizen and a person of color, or a U.S. citizen and an Indian American. The findings show that a single mailing seems to have little effect on the turnout rates of Indian American voters, although there is some evidence that multiple mailings may have had some effect. None of the identity appeals were especially effective at bolstering turnout.
Electoral Context: This experiment was conducted throughout the November 2004 national election.
Subject Population: The sample consisted of subjects identified on a list of registered voters in Queens County obtained from the New York City Board of Elections. From this list, Trevedi selected all those who had Hindu or Sikh Indian first and last names resulting in a sample size of 6,244 registered Hindu and Sikh Indian voters.
Randomization Procedure: Trivedi grouped individuals by households and randomly assigned households from the list to one of three treatment groups or the control group. The author stratified the sample by various factors including socioeconomic status, age, religion, prior voting history and other GOTV materials received.
Treatment: No mailings were sent to the control group. Trivedi sent each household one mailing one week prior to the November 2004 election, encouraging the voters therein to go to the polls on Election Day. The targeted identity of the mailings varied through different photographs and text. Other than the use of varying identifiers in two sentences on the mailing and the use of different photographs on each mailing, the postcards look exactly alike for all three treatment groups.
The first treatment group received a postcard mailing aimed at the identity of "U.S. citizen" with text that articulated the individual's duty as a U.S. citizen to participate in the electoral process. On the front of the postcard, there are four photographs of white individuals, thereby visually associating the term U.S. citizen with majority white America. The mailing received by the second treatment group builds onto the U.S. citizen identifier, addressing the individual as a "U.S. citizen and a person of color" with photographs of black, Latino, and Asian people, thereby visually associating "people of color" with significant racial and ethnic minority groups in the United States. The third treatment group received a mailing that also built onto the U.S. citizen identifier, adding the identity marker "Indian-American," and with photographs correspondingly of individuals who look as though they were ancestrally from the Indian subcontinent. All three mailings contain a paragraph that lists certain key issues on the political agenda that may have salience for Indian voters and other voters of color. These issue-based appeals stay constant for all three mailings.
Findings: Overall, receiving a mailing at all, without controlling for other variables, has a 1.1 percentage point effect on an individual's likelihood to vote. Although the effect is in the predicted direction, the estimate falls short of statistical significance (p > .05, one-tailed, taking intra-household clustering into account). This pattern of results is generally consistent with past literature, which states that mailings have anywhere from a 0 to 2 percentage point effect on a given population. The U.S. citizen appeal had the highest effect of the three treatments but remains statistically insignificant. It seems that modifying the identifier on a GOTV mailing has little effect to speak of, and identity appeals do not seem to provide a strong motivation to vote among Indian American voters in Queens.