Here's an explanation thread that goes beyond the obvious that people are complicated and can have identities that "we" consider incompatible.

(tagging @PollsAndVotes)

1/ https://twitter.com/PollsAndVotes/status/1284913145044258818

It's important to note that the CCES doesn't just weight by typical demographics. It weights by RELIGION particularly this born-again measure. It's not a major weight, but it is one of the religion measures it uses to weight.

2/

What's the comparison data set? It's not the census, obviously.

CCES uses @PewReligion Landscape study.

3/

Pew asks religion like this:

What is your religion (list with follow-ups for specific denominations)

And then, for Christians only: Are you a born-again or evangelical Christian or not?

4/

CCES reverses this for some reason.

Born-again then specific religion.

IMHO, this changes the meaning of the questions for a small but important set of respondents.

5/

If you ask denomination first, then born-again, you are asking only Christians if they id as born-again. They've told you they're Baptist, Catholic or whatever. So, it's asking about something else.

6/

But ask about born-again first, then the denom question will have a different meaning for some respondents.

7/

For example,

someone tells you that they're born-again and then you ask what they are and they say "nothing in particular" is saying that they're an evangelical who doesn't have a specific denomination.

8/

That is, they just said they were a Christian so this question is asking something different. Rather than asking about religion again (you just asked that!) it's asking about ethnicity or something.

9/

In Pew's survey they wouldn't have responded "nothing in particular" but likely "something else" because they are something (Christian)

10/

For minority religions that are also ethnicities (at least in the US) -- Jews, Muslims (to a degree), Hindus -- a denomination question after the born-again question might mean that they indicate their ethnicity when they otherwise may not have done so.

11/

There are some who would not have been asked by Pew the born-again questions bc they're not Christian. In the CCES they are asked.

The result cannot be fewer born-again Christians but more but then the weighting makes them equal.

12/

The result is that the Pew evangelicals are down-weighted (slightly) and these unusual "evangelicals" are up-weighted (slightly)

13/

One final note on Mormons, Catholics, and other Christians.

It's important to note that the "born again" language is not exclusive to "evangelical Protestants". It comes from the Gospel of John and many other Christians will answer that they are "born-again"

fin