Researchers Say More Americans Than Ever Before ‘Have No Religious Preference’

Another survey is telling us what we already know: The number of Americans who are not religious is on the rise. Researchers from University of California, Berkeley and Duke University went through the recently-released results of the biennial General Social Survey and found that

While some types of Americans identify with an organized religion less than others, Americans in almost every demographic group increasingly claim “no religion” since the trend began to accelerate in 1990.

This continues a trend of Americans disavowing a specific religious affiliation that began in the 1950s but has accelerated greatly since 1990.

Perhaps even more noteworthy: No matter how they sliced the population, every demographic group lost some religiosity since 1990 — many by huge margins:

Keep in mind that “atheism” isn’t necessarily growing at the same rate. We’re still a paltry 3.1% of the population. But the number of people who are giving up organized religion is higher than ever and that’s certainly something to be grateful about.

(No Cross No Crescent)

About Hemant Mehta

Hemant Mehta is the editor of Friendly Atheist, appears on the Atheist Voice channel on YouTube, and co-hosts the uniquely-named Friendly Atheist Podcast. You can read much more about him here.

  • http://www.youtube.com/user/GodVlogger?feature=mhee GodVlogger (on YouTube)

    Wow, look at that graph!
    Let’s see them try to deny *that* hockey-stick!

    • named

      I doubt they will deny anything.

      They’ll probably just do what they always do and condemn these people as those who are going to H-E-Double-Hockey-Sticks.

      • http://twitter.com/ylaenna M. Elaine

        They’ll claim it’s clear evidence of end times drawing near, the wheat being separated from the chaff. And also blame it on the gays, of course.

  • http://twitter.com/ylaenna M. Elaine

    Interesting how the percent change steadily increased with education through some college, then dropped at college grad, then went up again with an advanced degree. It seems to support what Chris Mooney says is the “Smart Idiot” effect.
    http://youtu.be/Jjzpm5mxQtE#t=10m50s

    • Randy

      The change column is only like that because the breakdown in 1990 was so weird. If you focus only on the 2012 column it breaks down exactly as you’d expect, the higher level of education, the less likely you are to be religious.

  • locohammerhead

    This is the internet and the age of information at work. It’s only going to speed up as younger generations age and older generations pass on.

    • abb3w

      That looks to be post hoc ergo propter hoc, and historically inaccurate. The shift of successive generations to increasing irreligion had fingerprints showing all the way back in the 1972 GSS data, when the internet was not particularly available to the public. Most of the current trend is because the basic religiosity of younger generations coming into survey age ranges is lower than the older generations that are dying out of the surveys.

      It’s possible some of the current waves of unaffiliation within cohorts or the Pew-documented increase in atheism among the unaffiliated are catalyzed by the internet. However, the former is small compared to the main logistic-curve shift against cohort; and the latter does not yet show clear fingerprints in the GSS samples.

      There’s other complications. In the GSS-2010 data, while higher irreligion was generally associated with higher internet usage, those without home internet at all were disproportionately likely for their generational cohort to be irreligious… which does not fit that conjecture. (I’ve not been able to extract that query from the GSS-2012. I’m also not sure the GSS-2010 results were statistically significant.)

      • locohammerhead

        You are right that it started earlier than mentioned but you completely missed my point. If you take a look at the graph at around 1987 (just around when the internet started to really take off) there is a very steep incline in non-religious affiliations ever since. Why is this? It’s a very simple explanation. Access to information. You can see non-religious affiliations were on the rise before then but moving at a very slow pace. They did not have access to the information we do now. All people had were their churches. So your claim that my post was post hoc ergo propter hoc is wrong and the graph proves it.

        • abb3w

          First off, “post hoc ergo propter hoc” is the Latin name for the questionable cause fallacy where claiming “Since that event followed this one, that event must have been caused by this one.” A graph of behavior against time at best only shows correlation, not causation.

          Second, I didn’t miss your point. I just think your claims are contradicted by a closer look at the data the chart is based on.

          Smacking the 2012 data about needs statistics software on your computer, but the 1972-2010 data can be tormented via a tool at Berkeley’s SDA. Differences in religiosity (RELITEN, which I recode to linearize as RELITENR) between generational cohorts (COHORT, which I usually recode by decades as CHRTDEC) are consistently larger than the changes of the cohorts over time (YEAR). The change in unaffiliation by cohorts fits fairly closely to a logistic curve, 27 year time constant, 2007 midpoint. This logistic curve against cohort is easily visible all the way back into the 1974 data (and also the 1972, if you look at RELIG). Given such a trend on cohort, most of the rise occurs as a result of that trend, and the changing cohort composition of the nation.

          There is a rise that begins in several cohorts circa the late 1980s. However, it’s more exactly the end of a decline in irreligiosity that had begun in 1972 or earlier. That looks to be more likely a combination of the rise and decline of the religious right, and possibly the boomers being religious long enough to raise kids.

          Contrariwise… using the fraction of each cohort for each sample GSS year, and the fraction of irreligious for each cohort expected by the logistic curve equation, one can compare the level of irreligion expected from such transition to the actual levels (if you’re willing to drag out a spreadsheet, or high-end statistics modeling package like SPSS). The early years of the GSS line up pretty well, but circa 1980 — the first survey year after the founding of the “Moral Majority” — the level of unaffiliation shows a depression, lasting through to… 1994, about when Septemeber Never Ended.

          Nohow, there’s a surge of “not very” religious in those cohorts circa that period, which fades out. You might give credit to the internet for helping the less religious overcome their fears of the religious right, but not for the underlying trend.

  • baal

    Note to Republicans, if tying yourself to religion is costing you votes, maybe you’d like to cut those strings to have an easier time at the ballot box? You wouldn’t have to use election fraud nearly as much and you’d be able to ditch Todd Akin-like canidates.

  • Frank

    Religious beliefs aren’t a “preference”, they are, well, beliefs. I don’t even know what this question is asking about, so I’m not sure how useful it is.

    • Ibis3

      How about thinking of it as how people prefer to identify themselves. If you’re given a choice: Catholic, Protestant, Jewish, Other, or None, which do you prefer as a label.

    • abb3w

      The exact wording is “104. What is your religious preference? Is it Protestant, Catholic, Jewish, some other religion, or no religion?” Protestants are also asked what specific denomination by the GSS.

      There’s a classic silly song about this. (In a twist of fate, Jerry Brown is again California’s Governor.)

  • Myke

    Interesting how the authors tried to downplay the massive growth of atheism since 1991, when in fact, our numbers have almost doubled, from an estimated 5.54 million in 1991 to over 9.69 million in 2012. “In 2012 only 3 percent of Americans chose the first answer ‘I do not believe in God.’ That was an insignificant change from the 2 percent who chose that response in 1991″ — Insignificant?!? That doesn’t apply to adding 4 million atheists, which is more than the population of Los Angeles proper.

    • abb3w

      Yes, “insignificant” — in the technically precise sense that the change is still too small to detect at the statistical 95% confidence level of significance in the GSS samples of N≈2000 from the total US population in each of the two years.

      It’s also very likely a non-trivial fraction of the people who self-identify as “Atheists” (say, to the Pew Research analog of the RELIG question) answer “I don’t know whether there is a God and I don’t believe there is any way to find out” as nearest to their beliefs rather than simply “I don’t believe in God“, and thus are counted for this as “Agnostic” — which has also increased considerably.

  • http://twitter.com/nicoleintrovert Nicole Introvert

    A crap ton of these people are still theists. This is the position many of my friends take. Or they just don’t care… a/theism just isn’t a part of their lives one way or the other.

    • abb3w

      Yep; and the PDF report notes that “Nones” and “Atheists” are not the same groups. The GSS results line up pretty well with the Pew Forum’s data; Pew has put up several studies about the nuances within the unaffiliated.

  • http://twitter.com/porlob Patrick Orlob

    Interesting that the massive uptick starts in 1995/96, right around the time that personal internet access started to become commonplace. I would not be surprised to learn that was a — or THE — major factor here. Anecdotally, I know of plenty of people who left religion because of the unprecedented access to information that the Internet brings.

  • abb3w

    The trend appears to remain that unaffiliation against birth cohort is on a logistic curve, with circa 2007 midpoint and 27 year time constant. Controlling for cohorts, there does appear to be a slight upswing in the number of unaffiliated and downswing in the strongly religious from that base trend; that may be a “social pendulum”, swinging back from the 1980s heyday of the religious right, which may itself have been swinging back from the 1950s or 1960s — though the GSS doesn’t go back far enough for any confidence about that.

  • cornell

    Darn, now I guess this means that my religion is automatically false. Oh well


CLOSE | X

HIDE | X