Pew Research Center illustration.

Over the years, Pew Research Center has asked Americans about the morality of both “homosexuality” and “homosexual behavior.”

Some respondents might be expected to answer these two questions differently because their religious traditions differentiate between identifying as homosexual and engaging in homosexual acts. The Catechism of the Catholic Church, for example, condemns unjust discrimination against people who are lesbian or gay, saying they “must be accepted with respect, compassion and sensitivity.” But it also lists homosexual behavior among the “sins gravely contrary to chastity,” placing it in the same category as masturbation, sex outside of marriage and pornography.

Does this difference in wording affect Americans’ responses to questions about the morality of same-sex relationships? We designed a survey experiment to find out.

How we tested our question wording about homosexuality

In a May 2025 survey, we asked 8,937 U.S. adults on our American Trends Panel whether they view each of several behaviors as morally acceptable, morally wrong or not a moral issue. We randomly assigned each respondent to receive one of two questions about homosexuality:

• “Do you personally believe that homosexuality is morally acceptable or morally wrong, or is it not a moral issue?”
• “Do you personally believe that homosexual behavior is morally acceptable or morally wrong, or is it not a moral issue?”

Randomly assigning respondents in this way allowed us to determine whether the question wording affected their responses.

Does the wording difference affect responses?

Americans overall make no distinction between the morality of homosexuality and homosexual behavior.

Similar shares of U.S. adults say homosexuality (39%) and homosexual behavior (41%) are morally wrong; the 2 percentage point difference is not statistically significant.

This pattern is broadly consistent, with no statistically significant differences across religious and demographic subgroups. For instance, there is no statistical difference between the share of White evangelical Protestants who say homosexuality is morally wrong (72%) and the share who say the same about homosexual behavior (76%).

Nor does the question wording make a significant difference among U.S. Catholics (34% vs. 37%). The larger difference among Hispanic Catholics specifically (11 points) is also not statistically significant, given the relatively small sample of Hispanic Catholics in the survey.

In addition, there are no significant differences between the overall shares of U.S. adults who say homosexuality and homosexual behavior are morally acceptable or not a moral issue.

However, among Catholics, 24% say homosexuality is morally acceptable, while 18% say the same about homosexual behavior. And among Hispanic Catholics, 25% say homosexuality is morally acceptable, compared with 13% who say this about homosexual behavior. While these differences are not statistically significant at a 95% confidence level, they are in the direction one would expect – greater acceptance of homosexuality than of homosexual behavior – if Catholics are attuned to the nuances of the church’s teachings on this topic.

Given the sample sizes of Catholics and particularly of Hispanic Catholics in our survey, we do not have the statistical power to conclude that the question wording makes no difference for these groups. (A total of 1,666 Catholics completed the survey: 832 received the question about homosexuality and 834 received the question about homosexual behavior question. This includes 193 Hispanic Catholics who were asked about homosexuality and 180 Hispanic Catholics who were asked about homosexual behavior.)

How should researchers ask about the morality of homosexuality in future surveys?

These results suggest that when examining how the public responds to questions about the morality of same-sex relationships, Americans give similar answers to questions about the morality of “homosexuality” and “homosexual behavior.”

However, survey researchers should bear in mind that the different question wordings may elicit different responses from some subgroups of the population, such as Hispanic Catholics.

Both wordings – asking about homosexuality or homosexual behavior – may sound outdated. Everyday language on these topics has been evolving, so some researchers might be inclined to ask, for example, about the morality of “having sex” or “a sexual relationship” with “a person of the same sex.”

This experiment cannot speak to whether such question wordings would yield results comparable to the historical survey trends on questions about homosexuality or homosexual behavior. If measuring long-term trends is a high priority, more experimentation is warranted before making changes to these questions.