Pew Forum's U.S. Religious Landscape Survey
In 2008, the Pew Research Center conducted a survey on religious affiliation in the United States. The survey was quite thorough, interviewing 35,000 adults through the country, in both English and Spanish, and was conducted by a well-respected organization. The study itself had few surprises, mostly reaffirming what people already know: Christianity is by far the largest religion in the U.S., with 78.4% of the adult population identifying as such, with all other religions and the non-religious far, far behind. The second largest group was the "unaffiliated," but this is too vague of a classification to mean much in itself. One aspect of the survey that was publicized by some religious groups on social media was a separate analysis by Georgetown University's Center for Applied Research in the Apostolate (CARA), which took the form of a graph showing retention rates of numerous religious affiliations. It showed the percentage of people raised in a faith (or lack thereof) who retained that same ideology as an adult. The group with the highest retention were Hindus, at 84%, and the lowest were atheists, at 30%. The overall results were as follows:
- Hindu: 84%
- Jewish: 76%
- Muslim: 76%
- Greek Orthodox: 73%
- Mormon: 70%
- Catholic: 68%
- Baptist: 60%
- Lutheran: 59%
- Pentecostal: 50%
- Buddhist: 49%
- Methodist: 46%
- Anglican/Episcopal: 45%
- Non-denomination Protestant: 44%
- Reformed: 42%
- Presbyterian: 41%
- Nones (nothing in particular): 38%
- Jehovah's Witness: 37%
- Congregationalist: 37%
- Holiness: 32%
- Atheists: 30%
Preach to the choir Religion |
![]() |
Crux of the matter |
Speak of the devil |
|
An act of faith |
v - t - e |
At first glance, this does seem damning for atheists, making this already small minority (just 1.6% of the population, according to the survey) appear doomed in the near future. However, further analysis of the statistics given by both Pew and CARA show quite a different trend.
Before we begin, it is worth noting that, as anyone familiar with statistics knows, the smaller the group, the more likely it is to be an outlier, sitting towards one end of the spectrum, while larger groups are more likely to trend towards the mean. This is no exception. The three groups with the highest retention, Hindu, Jewish, Muslim, were .4%, 1.9%, and .3% of the population, respectively[1] (all figures in this paragraph refer to the religion of respondents' childhood, not current religion). The largest of those, Jewish, is often viewed as an ethnicity as much as a religion, and therefore something that cannot be changed, making it something of an anomaly as well. On the other end of the spectrum we have atheists, Holiness, and Congregationalists, with .5%, .8%, and .8%. Sitting towards the middle (with retention rates spanning from extremes of 30% to 84%, the middle being approximately 57%) are the larger groups: Lutheran (59% retention, 5.5% of the population), Baptist (60% retention, 20.9% of the population), and Pentecostal (50% retention, 3.9% of the population). Catholics, the largest group at 31.4% of the population, are a little higher (68% retention), but still relatively close to the center. So out the outset, we already see what one would expect: smaller groups on the extremes, and larger groups more huddled in the middle.
Most of the numbers given in the graph appear to be accurate, as reported by Pew (a generally reliable organization), although a few are off by a percentage point, and some figures, including those for atheists, aren't published in the Pew study[2]. These were apparently culled from raw data supplied to CARA[3]. They should perhaps then be taken with a grain of salt, as CARA is religiously affiliated, but we'll accept them as is. What's missing from the graph is both the overall numbers, as well as what the people who left their childhood faith embraced in their adulthood. Fortunately, Pew supplied some of this information for us.
We'll start the analysis with the Catholics, as at least one version of the graph circulating on social media highlights Catholicism's relatively positive numbers, and their 68% retention rate is the highest of the major Christian groups in the U.S. (beat only by Mormons and Greek Orthodox, who are both very small groups). According to the survey, 31.4% of respondents were raised Catholic[4]. With the U.S. population in 2008 being roughly 304 million, this is roughly 95 million people. As we've seen, 68% remain Catholic as an adult, but 14% of them convert to no religion[5]. That's over 13 million Catholics who have since become non-religious. The survey does not identify how many became atheists, so we can only speculate. Regardless, the distinction between atheist, agnostic, and general "non-religious" has always been a hazy line. Somewhere around 10 to 15% of the unaffiliated are atheists (depending on whether "religious unaffiliated" is considered "religious" or "non-religious"), so if these numbers hold, that's between 1 and 2 million Catholics who have stopped believing in God altogether. Looking at the Protestants, the numbers are even greater. Looking at the same charts, we see that 53.9% of the population was raised Protestant (all denominations), which is roughly 164 million people. Of those, 13% converted to no religion, for a total of over 21 million. Again, if only 10% of those non-religious ex-Protestants have become full-blown atheists, that's roughly 2 million people.
But what about all those atheists who have converted to something else? As we've stated, Pew does not supply the figures for the conversion of atheists, but fortunately for us, CARA does. It is because initial atheist numbers are so small that this alleged 70% loss doesn't mean all that much, especially when we further explore the figures. According to Pew, just .5% of the population was raised atheist, which is about 1.5 million people. According to CARA's analysis of Pew's data, of those 1.5 million, 30% are now affiliated with a Protestant denomination, 10% are Catholic, 2% are Jewish, 1% are Mormon, and 1% are Pagan[6]. So that's approximately half a million childhood atheists who have since become Protestant, compared to several million ex-Protestant atheists, and around 150,000 ex-atheist Catholics, compared to 13 million non-religious ex-Catholics.
You also may have observed that with atheists retaining only 30% of their childhood members, and an accumulated 44% converting to other specific religions, we're left with 26% unaccounted for. Where exactly this quarter of the group falls we can't say exactly. It seems unlikely they fall into the other unmentioned religions (Muslim, Hindu, Buddhist, etc.) as these have very few adherents in the U.S. (none of those being more than .7% of the population), and it seems very unlikely that CARA would have neglected to mention it if there were large numbers of such converts. So we're left to assume that this mystery group fall into the "unaffiliated category", or the "nones". While the groups who circulated the graph of retention rates would certainly love to create the impression that 70% of atheists eventually find Christ, this is far from the case. Well more than half seem to retain their lack of religious conviction, and going from "atheist" to something like "agnostic" is more of a lateral transfer than anything else. So 56% of those raised as atheists remain without religion, and only about 40% embrace some form of Christianity, totaling maybe some 700,000 converted atheists, while Christians have lost something like 35 million to the ranks of the non-religious. (It's also slightly interesting to note that of ex-atheist converts, twice as many have converted to Judaism, which discourages conversion, as to Mormonism, who are constantly actively seeking converts.)
Finally, another problem arises when one tries to differentiate between being raised atheist, or simply non-religious. Being raised without any religious instruction at all could be called an atheist upbringing, or merely a secular one. Is someone who later converted to Christianity more likely to describe such an upbringing as "atheist" than an agnostic would? Perhaps. We can't say without further research, but there is certainly significant potential for selection bias when designating an atheist childhood in surveys such as this.