When discussing Latino voters, one question I often receive references Latino Catholicism. The question is basically: since Latinos are Catholics, and Catholics are anti-abortion, wouldn’t that make Latino Catholics more conservative and Republican?
No. This logic applies strongly to white voters but is far less descriptive of the Latino social setting. Just under half of all Latinos report some level of active Catholicism, while about a 23% identify as Protestant, including 15% Evangelical and 8% as Protestant non-Evangelicals. Among these two groups, Catholics are the more liberal, demonstrating the danger of extending findings and expectations regarding white voters to voters of color without good data.
In our recent study for UnidosUS, we interviewed 3000 Latino registered voters across the U.S. with oversamples in states with larger Latino populations, and battleground congressional districts. While Latinos and Catholics hold many opinions in common, there is a clear and discernible trend showing Catholics as significantly more liberal and more disenchanted with Trump than their evangelical counterparts.
Let’s start with the 2024 vote and potential regrets. Recalling 2024, Catholics report far higher Democratic votes than Evangelicals (60-34% vs 47-47%). When we asked voters if they would vote for the same candidate again, the results among self-identified Trump 24 voters was striking. Nationally, 25% of Latino Trump voters would not vote that way again. When we look at this by religiosity, 21% of Protestants indicate they would not vote for Trump again, compared with 27% of Latino Catholics. To the extent Latinos are moving away from Trump, Catholics are moving more quickly. By comparison, vote bleed away from Vice President Harris is 4% among Catholics and 7% among Protestants. Harris voters are by and large still in her column.
Presidential approval for Trump demonstrates the same pattern. Among Catholics, President Trump is underwater by 41-points (28 approve, 69 disapprove). These numbers are ten-points better among Protestants at both ends (38% approve, 59% disapprove). In fairness, Protestant Latinos are still net -21 on Trump, and far off their self-reported 2024 vote. Views about the parties in Congress more or less mimic this pattern.
On many surveys, we inquire about party reputations – that is, whether the voter thinks each party cares for them or is hostile. Among Catholics, 61% thought Democrats cared a great deal, 30% thought they didn’t care too much, and 9% found the Democrats hostile. Among Protestants, Democrats are viewed somewhat less favorably with 52% believing they care a lot, 40% saying ‘not too much,” and 8% finding them hostile. While the perceptions of hostility are about the same, there is a 10-point step down in Democratic favorability among Protestants.
Looking forward to this November and vote intention on the generic ballot, among Catholics the Democrats are +33 (58-25%), while among Protestants, the Democrats are +5 (43-38%). While there is considerable undecided among both groups, this is a rather large difference in vote intention and again suggests that the GOP bleed since the 2024 election has been far greater among Catholics than Evangelicals.
I have focused here on differences which, with the exception of 2026 vote intention, are more a matter of degree rather than direction. Ironically, the issue responses are very similar. When we ask which issues respondents most want to see addressed, both in terms of which issues matter and which aspects of each issue the respondent is most concerned about, each response is slightly (very slightly) more conservative or Republican for Protestants but overall the issue beliefs and policy preferences look pretty similar.
The evolution of religious identity among Latinos may signal a potential shift, but such a shift would be identity driven, rather than issue driven. And, with the “other” category also growing, it is too early to tell whether faith-based identity claims will provide a lasting challenge to Democratic hegemony among Latino voters.