As the United States heads toward the November election, Republican and Democratic candidates will be spending hundreds of millions of dollars courting Latino voters across the country. Latinos voted in record numbers in 2016, mostly against Trump, and are likely to be crucial in 2020.
So what works best at encouraging them to vote?
Here’s what we found in recently published research: English-language mailers work better than bilingual ones at getting Latino voters to the polls, as we’ll explain.
Why does the language matter?
Past research has found that Latinos are mobilized when encouraged to vote in their preferred language. When that’s done through ads on Spanish- or English-language television or radio, it’s easy to match the audience’s language preferences — because the audience has already chosen which language they prefer when they tuned in. On the phone or at the door, bilingual speakers can quickly switch to the voter’s preferred language.
But how do you match languages by mail? Despite the large databases on voters used in today’s campaigns, it is rare to find data on language preferences.
When candidates and campaigns reach out with bilingual or Spanish materials, they’re trying to let potential Latino voters know they’re inclusive and culturally sensitive. Strategists might imagine that Latino voters would see bilingual mailers as more culturally competent, especially for those who feel their fate is linked with that of other Latinos or are less acculturated to the English-speaking mainstream.
On the other hand, Latinos might dislike bilingual communication if they believe their citizenship or legitimacy is being challenged — for instance, if they find it insulting that someone assumes they might not understand English, or if they perceive the bilingual message as pandering or inauthentic. And an all-English mailer is shorter and less visually complex; it may be more effective just for that reason, or because the recipient sees it as more “official.”
Here’s how we did our research
To find out whether bilingual or English-language letters encouraging recipients to vote worked best at getting them to the polls, we conducted three experiments — one each in New Jersey and Virginia elections for the state legislature in 2015, and another in North Carolina elections in 2016, for the higher-profile general election for president, governor, senator and all the way down the ballot. We ran each in partnership with a nonpartisan civic organization in that state.
The tests were fairly large: We sent mail to 54,000 registered voters in New Jersey, 38,000 registered voters in Virginia and 123,000 (two letters each to 61,500 registered voters) in North Carolina. First, we used commercial data to identify Latino registered voters, using a predictive algorithm that uses surnames and Zip codes. This approach reflects the limits and challenges to finding Latinos, especially those who might prefer Spanish, in voter databases. In 2015, we included only registered Latino voters who sometimes vote; in 2016, we also included registered voters who had never voted before. Voters who were in the pool of targets for each state were then randomly selected to get mail in English, bilingual mail or no mail at all.
Each letter used what social scientists call “social pressure treatment.” It began by thanking the recipient for the last time they voted, mentioning the year, or for registering to vote. It listed several past elections, noting whether the recipient had cast a ballot. It went on to urge the recipient to vote this next time when “many people like you will be voting,” and offered information about voting hours and poll locations. In New Jersey and Virginia, this included links to the official state election office website. The North Carolina 2016 letter also encouraged early voting; those letters included a link to the local County Board of Elections and their phone number, as well as a list of one-stop early voting locations and hours.
Social pressure treatments gently remind recipients that voting is a matter of public record, and that their past (and future) voting behavior is being observed. Because voting is a social norm — people know that they are “supposed” to vote — they’re more likely to do so if they have been reminded that others will know about it.
Half the recipients got only a letter in English; the other half received a letter that had, below the English, a Spanish translation in italics. After each election, we checked administrative records from state election officials to confirm whether each individual on our lists had voted.
The letter in English increased turnout
The English letters were the big winners, by far — increasing turnout more than the bilingual letters, from 35 percent to 70 percent more.
In both New Jersey and Virginia, getting any letter made someone more likely to vote than those who did not get a letter. But those who got English-language mailings were much more likely to vote than those who got the bilingual ones.
In New Jersey, 11.5 percent of those who received no letter voted; of those who got a bilingual letter, 13.7 percent voted; and of those who got an English-language letter, 14.5 percent did. All these results are statistically significant.
In Virginia, of those who got no letter, 11.5 percent voted; of those who got the bilingual letter, 12.9 percent; and of those who got the English-language letter, 13.4 percent. Again, all these results are statistically significant, although the difference between the English and bilingual mailings was only marginally so.
In North Carolina in 2016, of those who got no letter, 48.7 percent voted; of those who got the bilingual letters, 49.4 percent; and of those who got the English language letters, 49.9 percent. In this case, while the English letters increased turnout in a statistically significant way, that wasn’t true for the bilingual letters.
Don’t assume two is better than one for languages
These differences may seem small. But when elections are closely contested, small differences in voter turnout can have big effects.
To be sure, Latinos are a relatively small proportion of the electorate in these three states. Bilingual mailings may be more effective elsewhere or with better tools to identify Latinos who prefer Spanish.
Christopher B. Mann is assistant professor of political science at Skidmore College. Melissa R. Michelson is dean of Arts & Sciences and professor of political science at Menlo College. Her most recent book is “Transforming Prejudice: Identity, Fear, and Transgender Rights” (Oxford University Press, 2020). Matt Davis is vice president of civic engagement at League of Conservation Voters Education Fund.