Firstly, this is a poll. Remember, polls are bullshit.
Secondly, this is the Public Religion Research Institute. Made possible by a grant from the Unitarian church. (Not that there’s anything wrong with that, just that that’s the case)
Thirdly, follow the link to the poll to see breakdowns like, “ Strong majorities of Americans — including most people of faith — support nondiscrimination protections for LGBTQ individuals, but overall support has declined.” Do you use the term “people of faith”? No? Why or why not? That’s the third point against.
Fourthly, it’s a relatively large sample size, 22k, compared to the 800-2000 we usually see, using the Ipsos KnowledgePanel. Ostensibly a good thing. Link is at the bottom of the survey page. KP, for short, is an online poll. People get a random letter in the physical, postal mail. The letter says “Hi (your name here) we’re a super respectable polling agency who’d like to make money off your opinions” and includes a special seekrit password to “let” you sign up. So all the respondents did that. Would you do that?
Then, weeks or months (or years?) later you (as a KnowledgePanel Invitee extraordinaire) get a random email that says “go online and give us your opinion on ‘matters of faith’” (i’m just making that up, but they could have used that language)
Then, Fifthly, people went online to share their real honest and true thoughts about LGBTQIA+ protections and other matters “of faith”. Does being online skew people’s opinions? You’d always answer online as you would face-to-face wouldn’t you?