Spanish-language Radio Advertisements and Latinx Voter Turnout in the 2006 Congressional Elections
Latinx individuals in the United States vote at lower rates than other ethnic and racial groups, and political campaigns target Latinx audiences less frequently than other ethnic groups. Many Latinx voters in the US primarily speak Spanish and may not be reached by general, English-language campaign outreach. Researchers tested the impact of non-partisan Spanish-language radio campaigns on Latinx voter turnout in the November 2006 congressional elections. These radio advertisements significantly increased Latinx voting rates.
In the United States, the voting rate among Latinx individuals, who constitute 16.3 percent of the total population, is lower than that of other ethnic and racial groups.1 Although the Latinx population is growing rapidly, research suggests that political campaigns often neglect Latinx voters relative to other ethnic and racial groups. Efforts to increase the voting rate among the broader population may not be effective for this segment of the population, as language barriers could make it difficult for political campaigns to reach Latinx voters. Given that radio exposure is higher among Latinx individuals than other racial or ethnic groups, Spanish-language radio advertisements could be an effective way to increase Latinx voter turnout. While previous research has found that of radio campaigns have had no significant effect on voter turnout, there is very little evidence on the impact of targeted, Spanish-language radio advertisements on Latinx voter turnout.
Context of the evaluation
In the November 2006 congressional elections in the United States, 32 percent of eligible Latinx voters voted, compared to 52 percent of non-Hispanic whites and 41 percent of Black voters. In addition to these comparatively low voting rates, Latinx individuals also volunteer in campaigns, contact elected officials, and contribute to candidates less frequently than other groups. Their lower participation rates could be one explanation for why political campaigns target Latinx voters less frequently.
Many Latinx people in the US primarily speak Spanish and rely heavily on Spanish-language media. Only 28 percent of those interviewed in the 2006 Latino National Survey indicated that they could hold a conversation in English “very” or “pretty” well. Moreover, the 2000 US census indicated that 28.1 million Americans spoke Spanish at home. Consequently, the Latinx population relies heavily on Spanish-language media. In 2004, two-thirds of Latinx adults reported receiving at least some news from Spanish-language sources, with Spanish-language radio being particularly popular, only surpassed in use by television.
Details of the intervention
Researchers designed a randomized evaluation to test the impact of Spanish-language radio advertisements on the Latinx voting rate in the November 2006 general elections. They partnered with a political consulting firm to record and produce nonpartisan 60-second radio advertisements to encourage listeners to vote, providing a list of important issues in the upcoming elections, and the names and party affiliations of the candidates.
During the week preceding the elections, the advertisements aired in 36 congressional districts randomly selected from a group of 206 congressional districts. This sample included all congressional races where an incumbent was running for reelection against at least one challenger. Researchers only considered congressional districts where the race included candidates from the two major political parties and excluded very competitive races and districts where data on prior Latinx voting rates was unavailable. They also excluded congressional districts near Los Angeles and New York City due to the high cost of advertising in these areas.
Researchers purchased airtime from stations that reached a Spanish-speaking audience. The amount of advertising time purchased varied in each district. Districts were exposed to 50, 75, or 100 gross ratings points (GRPs) of advertising, depending on each individual district’s cost per GRP. GRPs are used to measure the size of an advertisement’s audience and factor in both the reach (the percentage of the target population that has heard the ad at least once) and frequency (the number of times the audience is hears the ad) of an advertisement. The stations played the ads during peak listening times during morning and evening rush hour, as well as during the day and on the weekend.
Following the elections, researchers examined voter registries and obtained verified voting histories for Latinx voters in each district to create district-level voter turnout percentages.
Results and policy lessons
The Spanish-language radio advertisements significantly increased the Latinx voting rate in the 36 congressional districts where they aired. Researchers found that purchasing 100 GRPs (the maximum amount purchased per district) of advertising raised Latinx voter turnout in the 2006 congressional elections by 4.3 percentage points on average. The overall Latinx voting rate across all 206 congressional districts in the sample was 34.5 percent.
The presence of these radio ads could potentially increase the voting rate of other ethnic and racial groups, since any person with a radio, Latinx or not, could listen to the ads. However, media research shows that non-Latinx audiences listen to very little Spanish-language radio. Researchers found no relationship between the ads and voter turnout among other ethnic groups, confirming that only the target audience responded to the ads.
The campaign spent approximately $9 in advertising for each additional Latinx voter. Compared to door-to-door, mail, and phone campaigns, which previous research has shown must spend $30, $60, and $100, respectively, per additional voter, the radio ads were a cost-effective means of increasing voter turnout.
Panagopoulos, Costas, and Donald P. Green. 2011. "Spanish-language Radio Advertisements and Latino Voter Turnout in the 2006 Congressional Elections: Field Experimental Evidence." Political Research Quarterly 64(3): 588-599.