Nineteen Sixty-four is a research blog for the Center for Applied Research in the Apostolate (CARA) at Georgetown University edited by Mark M. Gray. CARA is a non-profit research center that conducts social scientific studies about the Catholic Church. Founded in 1964, CARA has three major dimensions to its mission: to increase the Catholic Church's self understanding; to serve the applied research needs of Church decision-makers; and to advance scholarly research on religion, particularly Catholicism. Follow CARA on Twitter at: caracatholic.


The Island of Misfit Polls

There you go again (1, 2, 3)... According to the Pew Research Center, "Catholics appear to be declining both as a percentage of the population and in absolute numbers" in the United States. But are Catholics disappearing anywhere else other than within in the walls of the Pew Research Center?

As shown in the figure above, the self-identified Catholic affiliation percentage since 2010 has mostly varied between 21% and 26% among surveys conducted by Gallup, Pew, PRRI (data are only available for analysis from 2010 to 2013), and the General Social Survey (GSS). The average of all these polls is 23.2%. This is generally consistent with a trend that began in the late 1940s in Gallup's surveys and has persisted through the GSS series that began in the early 1970s. The only series on the figure above that shows a downward trend (i.e., the dashed lines) is Pew's. The current study estimates that 20.8% of U.S. adults are Catholic. This is down from 23.9% in a similar study conducted by Pew in 2007.

Pew is taking the difference between these two survey estimates quite literally; so much so that they have attempted to do a population trend analysis with just these two data points. It takes a special kind of hubris to treat survey data as if it were a census and declare a trend with an N of 2 (...knowing that this result was not generally evident in other polling, which is noted well after the declaration on page 115 in Appendix C of the report). This will either turn out to be a brilliant claim ("we said it first") or a bit of a blemish for survey research. Time will tell and to be honest it already has as the 2014 GSS came out of the field after Pew and that study estimated 25.4% of U.S. adults are Catholic. As the principal investigator for the GSS noted in the Wall Street Journal today, "There’s no hint of any decline."

I've heard Pew researchers claim that the question wording they use allows people to be honest and more easily admit they have no religious affiliation (i.e., reducing social desirability bias). But I believe PRRI uses the same question and I don't see the same trend in their data that is evident in Pew's surveys (PRRI's American Values Atlas estimates that 23% of adults were Catholic in 2014). I don't think it's the question wording.

The lower than average Catholic estimate in Pew's study could be a reflection of some sampling issue. As noted on page 71 of Pew's report, only 48% of Latino respondents self identified as Catholic in the study. Among the surveys (using bilingual or English and Spanish interviewers) shown in the figure above, the typical poll estimates a majority of Hispanic or Latino adults self-identify as Catholic. Accurately sampling and surveying Spanish-speaking Hispanics/Latinos in the United States is one of the more challenging tasks survey researchers have. CARA's meta-analysis of recent national surveys indicates this can often be the source of a survey estimate for Catholic affiliation "falling low" when aggregated with other studies.

As I often tell my students at Georgetown, surveys are blurry portraits of reality. Treat them as such. There is always more to worry about than the margin of sampling error. Confidence intervals still exist even when you interview 35,000 people. And beyond that there are many potential error components that have nothing to do with chance. I have no doubt that the number of Americans who are religiously unaffiliated is rising. There are pieces of data that indicate to me that the Catholic affiliation percentage in the U.S. will fall in the future ( the same time, the Catholic population will likely continue to grow). As a scientist I will always follow the data. But I am rarely convinced by a finding that is inconsistent with what most others are registering. I've read Pew's report. I'm generally a fan of their work. But I still feel I have no credible evidence that the Catholic population is declining in the United States. It is possible but not very likely. In other words, statistically speaking, don't bet on it. 

Search This Blog


Blog Archive

© 2009-2015 CARA, Mark M. Gray. Background image courtesy of muohace_dc.