Inequality and Heterogeneity Strike Again

One of the important characteristics of mathematical sociology is that it can allow us to make predictions about how the world should work without a lot of detailed substantive knowledge. Obviously, this advantage seems like a handicap to some scholars, but I think otherwise and a recent study suggests that I may be correct. I refer, of course, to a recent survey conducted by the Pew Research Center that has been getting a lot of attention. It finds that Atheists and Agnostics have more religious knowledge than any other group in the U.S.:

Atheists and agnostics, Jews and Mormons are among the highest-scoring groups on a new survey of religious knowledge, outperforming evangelical Protestants, mainline Protestants and Catholics on questions about the core teachings, history and leading figures of major world religions.

On average, Americans correctly answer 16 of the 32 religious knowledge questions on the survey by the Pew Research Center’s Forum on Religion & Public Life. Atheists and agnostics average 20.9 correct answers. Jews and Mormons do about as well, averaging 20.5 and 20.3 correct answers, respectively.

Protestants as a whole average 16 correct answers; Catholics as a whole, 14.7. Atheists and agnostics, Jews and Mormons perform better than other groups on the survey even after controlling for differing levels of education.

Needless to say, this finding has journalists in a tizzy, perhaps especially because- as research by Edgell et al. recently demonstrated– atheists are not well-liked, often being perceived as elitist, snooty and amoral. Clearly, it would seem, such persons could not know about religion.

Table shamelessly borrowed from Pew

Even when you break the questions out by category, distinguishing knowledge of Christianity and the bible from knowledge of other world religions from knowledge of religion in U.S. public life, atheists and agnostics do as well or better than average in providing correct answers. And an appropriate question is, of course, why?

There are a number of theories, of course. Dave Silverman of American Atheists suggests in a recent New York Times article that the reason is because atheists and agnostics have spent a lot of time thinking about religion:

“I have heard many times that atheists know more about religion than religious people,” Mr. Silverman said. “Atheism is an effect of that knowledge, not a lack of knowledge. I gave a Bible to my daughter. That’s how you make atheists.”

And Greg Smith, one of the study authors, makes a similar claim to CNN:

“Very few people say that they were raised as atheists and agnostics,” he explains.

About three out of four were raised as Christians, he says.

“They were raised in a faith and have made a decision to identify themselves with groups that tend to be fairly unpopular,” atheists and agnostics, he says.

“That decision presupposes having given some thought to these things,” which is strongly linked with religious knowledge, he says.

These claims may be correct, but they leave behind a small problem: why is it that the religious knowledge of Jews and Mormons is nearly as high as atheists? Indeed, the difference is probably not statistically real and yet many, if not most, Jews and Mormons were probably born into their faiths. Can we come up with a common explanation for all of their high scores?

Indeed, we can, thanks to the work of a giant of mathematical sociology, Peter M. Blau. In 1977 Blau published his classic “Inequality and Heterogeneity: A Primitive Theory of Social Structure,” in which he spent close to 300 pages explaining how simple aspects of population distributions could be used to explain a wide variety of social behavior. If you’re not in the mood to read 300 pages, you might also consider reading his 1977 American Journal of Sociology article titled, “A Macrosociological Theory of Social Structure,” which is essentially the short version of his book. I will focus on the article since I expect most of you don’t have the book handy.

In any case, among other topics Blau spends a great deal of time considering what is logically implied by a disparity in the sizes of two groups in a given society, eventually leading to his first theorem (AJS article, Pg. 35), as well as three sub-theorems:

T1: In the relation between any two groups, the rate of intergroup associations of the smaller group exceeds that of the larger.
T1.1: The proportion intermarried (or having another exclusive assocation, as mutual best friends) in the smaller group exceeds that in the larger.
T1.2: The mean number of intergroup associates in the smaller group exceeds that in the larrger.
T1.3: The mean amount of time spent in intergroup associations in greater for the smaller than for the larger group.

Essentially, Blau is arguing that if we have two groups, A and B, and those groups associate with each other, the rates of association will be different. Say that group A contains 10 persons while group B contains 100. If there is one cross-group association, then one A and one B associate with each other. This single relation, however, means that 10% of the members of A know a B, while only 1% of members of B know an A. If we increase the number of relations to ten and distribute them evenly, then 10% of members of B would know an A, while 100% of members of A would know a B. Thus, barring some very odd circumstances (e.g. there are ten relations that all include the same A but connect to ten different Bs) we can be fairly confident that a larger percentage of As will have direct contact with Bs than vice versa.

But what does this mean for knowledge of religion? Simply this: much of what we know about others, particularly the beliefs of others, we learn through contact with those others. If you are a member of group A in the above, you therefore will have a difficult time not knowing someone in group B or, alternatively, someone in group A who knows someone in group B. In contrast, if you are in group B, it will be quite easy to not know anyone in group A, or to not know anyone who knows someone in group A. If you never come into contact with an A, you may never learn very much about what makes them an A and not a B. And likewise, the heightened knowledge of religion exhibited by atheists, agnostics, Jews, and Mormons may all be attributable to their relatively low representation in the population. As minorities- in some cases very small minorities- they cannot help but know quite a bit about the majority that surrounds them but, by the same token, that majority may have fewer opportunities to learn about the minorities.

Are there other possible factors? Of course! The explanations offered by Smith and Silverman may well be relevant factors. Likewise, the emphasis on missionary work among Mormons may explain some of their exposure to, and knowledge of, other faiths. Yet, without even considering these specific factors, we already have reason to believe- thanks to Blau and mathematical sociology- that minorities should know more about majorities than the reverse. In other words, we can likely explain quite a bit of the effect using the simple numbers of such persons in the population, without having to postulate some more nebulous effect of different beliefs.

Mathematical sociology is often abstract, but it is that abstractness that lends it such power.


Blau, Peter M. 1977. Inequality and Heterogeneity: A Primitive Theory of Social Structure. New York: The Free Press.

Blau, Peter M. 1977. “A Macrosociological Theory of Social Structure.” The American Journal of Sociology, 83(1): 26-54.

Edgell, Penny, Joseph Gerteis and Douglas Hartmann. 2006. “Atheists As ‘Other’: Moral Boundaries and Cultural Membership in American Society.” American Sociological Review, 71(2): 211-234.


8 Responses to Inequality and Heterogeneity Strike Again

  1. Michael Bishop says:

    Neat connection to Blau… I had read about the Pew research elsewhere, but not made that connection myself.

    Its no surprise that atheist groups promote the idea that religious knowledge causes atheism. Perhaps, but I imagine that atheists, being better educated, would have outscored the religious on a test of history or physics as well. Atheist groups would probably be almost as happy to hear that anyway.

  2. OK, I don’t get it.

    It seems like the Blau model (or for that matter, the DuBois concept of “double consciousness”) would explain why a little group would no more about a big group than vice versa. However I don’t see any reason why this implies that the minority should know more about the majority than the majority does about itself.

    I understand how if you know a lot about your own little group and a lot about the big group this implies more knowledge overall than knowing a lot about your own big group and nothing about the little group. But this doesn’t explain the results broken out by question topic. Note that atheists greatly outperform Catholics (and are comparable to non-evangelical Protestants) even on the questions specifically about Christianity. Likewise, the gap between Mormons and non-evangelical Nicene Christians is too big to explain by assuming that the Mormon knows the same amount as them about Martin Luther, the eucharist, etc, but also knows who Joseph Smith was.

    Basically, you’ve got to explain why a Mormon would know more about transubstantiation than a Catholic. It makes no sense to say “because a Mormon will know more Catholics than vice versa” because apparently the Mormon’s numerous Catholic friends are all too poorly catechized to explain anything.

    Speaking of religious ignorance, yesterday in my lecture of 150 undergrads not a single one of them had ever heard of the Jonathan Edwards sermon “Sinners in the Hands of an Angry God.” WTF do they teach in high school US history these days?

  3. mattbrashears says:

    Hey Gabe,

    Well there you go trying to ruin a good story!

    All kidding aside, though, I’m not sure I agree with you. I see your point, but it’s worth wondering whether religious content would tend to be transmitted equally in all interactions. In the case of a Christian-Christian interaction, the religious views are, or are presumed to be, similar and thus religion doesn’t inevitably become salient. In an Atheist-Christian or Jew-Christian interaction on the other hand, religion is far more likely to become salient. Aggregated over many interactions, this means that religion is frequently a salient issue for minority religions, and relatively infrequently a salient issue for majorities.

    It’s roughly comparable in concept to how racial minorities in the U.S. are often cognizant of the impacts of race, whereas whites can go through their lives blissfully unaware.

  4. Michael Bishop says:

    I’m certainly inclined to believe relative group-size doesn’t explain the entire difference. But its an interesting theory, and I wouldn’t be surprised if it plays a role. If I’m correct that atheists are more knowledgeable about non-religious topics that would be another clue that there was something more than group-size at play. Regardless, I think it profitable to consider Blau’s work in this context.

  5. i think the three of us are developing something of a consensus, which is that it’s really about salience.

    one thing that is likely to produce salience is being a member of a minority, especially a minority like Mormons that is stigmatized and whose relationship to the majority is actively contested. (Google gives 762,000 hits for “are mormons christians”). of course there are other paths to salience too, which could explain why evangelicals are pretty knowledgeable even though they are a big group. note that Catholicism is neither small nor salient (many cradle Catholics treat it as a life course temple religion) so that explains why so many Catholics answer questions with a rate only slightly better than guessing at random.

    however note that we’re talking about the meaning of being a minority. this is a bit more fuzzy and cognitive than the Blau thing, which was more about mere contact under a null of random sorting — a kind of social law of thermodynamics.

    btw, who was the person who stated the Blau model with the rather creepy example that as a higher proportion of professors will sleep with their students than vice versa?

    • Jim says:

      OK I just found this blog, so I hope I am not to late to the party.
      This discussion seems like it is screaming for a model of learning, or knowledge diffusion that uses salience of differences as well as the probability of interaction. I can see this model working to help explain innovation in society and connecting to work on decision making and innovation in teams.
      I would imagine that a very nice formal model could be developed. Is anyone aware of any work that has been done along these lines?


  6. jimi adams says:

    Coming a little late to this party, but somehow i missed this post in my feed-reader. In case you haven’t seen it there are some interesting contributions (some, obviously more useful than others) to interpreting these findings from a variety of religious scholars on the Immanent Frame – Doesn’t help tease out the discussion you all are having much, but i thought it was interesting related reading.

  7. […] Inequality and Heterogeneity Strike Again « Permutations "Atheists and Agnostics have more religious knowledge than any other group in the U.S." (tags: usa:religion) […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: