Last Monday, the Pew Forum on Religion & Public Life released the first analytical hunk of what it is calling the U.S. Religious Landscape Survey (not to be confused with earlier Pew-sponsored “American Religious Landscape and Political Attitudes” surveys). Thanks to the Pew name, a fab marketing strategy, and a really cool website, it has received a huge amount of media attention—despite the fact that it offers precious little in the way of new information about the American religious landscape, and some of what it purports to have found is actually, well, misleading.

As a good case in point, take what has been the big take-away for a lot of media: the increase in the number of Americans who are, in Pew language, religiously “unaffiliated.” (These are folks who academics usually refer to as Nones—because when asked what is their religion, they say “none.”) Pew found that 16 percent of Americans fall into this “unaffiliated” group. It also found that only seven percent of the population said they were raised unaffiliated. On Wednesday, when Brian Lehrer of WNYC asked Pew’s Greg Smith in a radio interview if that meant that the number of unaffiliated had doubled in, say, 10 years, Smith said, “It’s not that there was ever a particular point in time where seven percent of the public was unaffiliated. It’s that that’s how many adults today were raised unaffiliated.”
But that’s not true. Other studies, including not only the General Social Survey (conducted by the National Opinion Research Center at the University of Chicago) but also, more importantly, my colleague Barry Kosmin’s National Survey of Religious Identification (NSRI) showed precisely that in 1990. And his and Ariela Keysar’s American Religious Identification Survey (ARIS) showed that the number of Nones had doubled by 2001 to 14 percent. That was news. (By the way, NSRI’s 120,000 sample and ARIS’ 60,000 sample were far larger than the 35,000 surveyed by Pew. Both are very well known—indeed, have for the past six or seven years been cited in the Census Bureau’s Statistical Abstract of the United States.) Meanwhile, in 2004, Pew’s American Religious Landscape survey showed “unaffiliateds” at 16 percent. The point is: The real news of the Landscape Survey regarding the unaffiliateds —and the story that should have been written—is that since 9/11 their numbers have stabilized at about 15 percent. This segment of the population has not grown during this decade. The growth occurred during the 1990s.
Now, it should be noted that the actual report put out by Pew does make mention, in the introduction, of the existence of other research in this area, including ARIS. And there is a chart of GSS findings on the rise of the unaffiliateds. But the whole tenor of Pew publicity has been to hype the 16 percent unaffiliated as a “key finding,” as if the Survey had discovered something new.
Let’s consider another example: the case of the Catholics. According to the Landscape Survey, 31 percent of Americans say they were raised Catholic, but only 24 percent of Americans are Catholic today. (This comes from asking what your religion is today, and what it was when you were a child.) Now that suggests a huge drop-off in the proportion of Catholics in America. But in fact, there is exactly the same proportion of Catholics in America today as there was a generation ago. What’s happened? Well, a lot of non-Catholics have converted to the faith and immigration to America has been disproportionately Catholic (n.b. Latin America, the Philippines, Vietnam). But in its introduction, the Landscape Survey report only cites immigration: “These losses sould have been even more pronounced were it not for the offsetting impact of immigration.” The take-away is thus supposed to be that Catholics are losing ground. But in fact, the number of Catholics in America has risen in exact proportion to the growth of the general population. In other words, the story here should be: Catholics are holding their own.
Finally, there an oddity in the Muslim numbers. A year ago Pew released a smaller, targeted survey of Muslims that put their number at .6 percent of the population—about 2 million. Because that was less than a third of what national Muslim organizations like to claim, Pew caught a lot of flak—as indeed, the ARIS researchers caught when they came up with a Muslim percentage of .5. This is a very politically fraught subject, and so one naturally wants to know what this big new survey came up with. Surely that would be a “key finding.” But what does Pew do? It simply refers to its earlier, less comprehensive survey. Who’s kidding whom? What’s the real number guys? Might it be less than the .6 percent found earlier?
What lessons can be draw from all this? Well, one is that it is important to recognize just how important publicity is to Pew when it comes to the research products of what it calls its “Fact Tanks.” (Not “Think Tanks, mind you; the ones like the Pew Forum that are interested in just the facts, ma’am.) At the Pew Charitable Trusts, it is very, very important to be seen to having an impact, to be moving the needle. Not that Pew is alone in this, when it comes to religion surveys. The 2006 Baylor Religion Survey was similarly vectored toward maximum public impact. Mind you, I’m not against tooting one’s horn. But sometimes it’s mostly just toot. Beyond that, it’s worth keeping a sharp eye out for what’s being pushed out as news and what’s not. Finally, it’s not a good thing for institutions to act as if they are the only game in town. Yes, yes, the full report may make that sufficiently evident in the fine print. But research outfits, even operating inside the Beltway self-promotion world, need to hold themselves to academic standards of transparency and fair play.
As for what lies ahead, the word on the street is that the Landscape Survey has some real news in what lies ahead. One can hope for more honesty in packaging when that day comes.