A Public Religion Research Institute poll led to some headlines claiming that white evangelicals think they're discriminated against more than Muslims. They're not reading the data correctly.
Here's a sample of some of the headlines:
What the PRRI poll showed is that more white evangelicals believe Christians are discriminated against than believe that Muslims are discriminated against. The poll cannot tell us what percentage of white evangelicals think they're discriminated against more than Muslims because that's not the question the survey asked.
The actual survey question looked like this: "Just your impression, in the United States today, is there a lot of discrimination against [INSERT ITEMS; RANDOMIZE], or not? And is there a lot of discrimination against [INSERT NEXT ITEM], or not?"
The items inserted included several different groups, including "Muslims" and "Christians."
What PRRI did not ask was for respondents to compare the amount of discrimination against Muslims to the amount of discrimination against Christians, which is what the headline writers above presume to know.
It probably didn't help that PRRI itself also got the headline wrong on its own chart included in the report, writing, "White Evangelicals Perceive More Discrimination Against Christians Than Muslims." (Update: I spoke with PRRI's Daniel Cox after publication of this article. While we didn't agree completely on how to interpret this data, he agreed to change the chart headline to, "White Evangelicals More Likely to Perceive Discrimination Against Christians than Muslims.")
PRRI got it right, on the other hand, when describing the results in the actual report and when it tweeted the table and wrote, "Most (57%) white evangelicals say Christians face a lot of discrimination in US today; 44% say same of Muslims."
The report also does not say what percentage of white evangelicals think Christians are discriminated against but Muslims are not discriminated against. To know that answer, one would need a crosstab of white evangelical answers to Christian and Muslim discrimination.
That crosstab would tell you what percentage of white evangelicals fall into each of four groups:
1) think both Christians and Muslims are discriminated against,
2) think Christians are discriminated against but Muslims are not,
3) think Muslims are discriminated against but Christians are not, or
4) think neither Christians nor Muslims are discriminated against.
Many of the articles about that part of PRRI's poll assumed white evangelicals fall into category 2 without considering the other three options. Without a crosstab of those answers, we don't know what portion the 57 percent of white evangelicals who believe Christians are discriminated against do not think Muslims are discriminated against. It could be as few as 12 percent (57 minus 44) or as high as 56 percent (100 minus 44).
I contacted PRRI to find out what those crosstabs would show. PRRI told me they don't have the answer because they didn't ask all white evangelicals the same question.
Daniel Cox, research director for PRRI, explained Monday via email: "Unfortunately, we can't run the analysis you requested. On these types of survey we sometimes randomly divide the sample so not all respondents get every question. This enables us to ask considerably more questions than we would otherwise be able. One downside of doing this is that it's not possible to run an analysis on some pairs of questions. This happens to be the case of the question on discrimination against Muslims and discrimination against Christians. They were each asked of a different subset of respondents. It's always a trade off when we make these decisions."
Another important question is whether there are differences based upon religiosity. Are more highly religious white evangelicals any more or less likely to perceive discrimination against themselves or Muslims than less religious white evangelicals? Unfortunately, PRRI didn't include any measures of religiosity, Cox said.
It may be that a majority, or significant minority, of white evangelicals perceive more discrimination against themselves than against Muslims. That would certainly be a finding worthy of debate. But the PRRI poll does not answer that question.