Wednesday, February 27, 2008

Religious demographics from Pew: Mormons get it done at office, in bedroom

++Addition++GNXP's Razib Khan, the epitome of why I favor a merit system crudely along the lines of the EB-5 Visa program as opposed to a moratorium on immigration, takes notice.

Also, a commenter points out that a non-significant number of Hindu men in the US have spouses that are living back home. So the 71% of Hindus in the US who are married to other Hindus could theoretically be comprised of the 61% of the total Hindu population in the US that is male, with five out of six of them married to a spouse living overseas, and only one-third of Hindu women married to other Hindus. But 86% of Hindus surveyed were foreign-born, and over 90% of people of Indian descent (both genders) in the US have spouses that are also Indian. So the pickings are still slim, but maybe not that slim. And as Agnostic points out (via Razib), beyond the first generation, Indian women are not averse to marrying white men. The Manjula reference holds, since she was born and raised in India, although I didn't initially think it through with that in mind!


Pew Research just released a demographic report on religious affiliation in America. Due to a law passed in the mid-seventies, the US Census is unable to mandate information concerning religion be volunteered by the population. Consequently, much of the data available come from religious organizations themselves. That leads to a lot of apples-to-oranges comparisons and an incomplete picture.

What immediately caught my eye, in light of the surge of attention given to Mormonism, is how its adherents are both moderately affluent and fecund. I used 20 of Pew's classifications to build a self-contained linear equation predicting what percentage of each categories' population 'should' have three or more kids living with them.

The equation was created by running a regression on the percentage of each group earning $50,000 or more per year and the percentage of each group with 3 or more kids under the age of 18 currently living at home, adjusting the 3+ kids percentages for the portion of each group aged 65 and older (I presumed no one in this age group had any children living at home under the age of 18 and removed them from the 3+ kids percentages to avoid an age skewing as much as possible). The two variables inversely correlate at a firm .44. That is, by religious affiliation, as income increases, the number of children decreases.

Included are the sub-category figures for Protestants, Jews, "Other faiths", and "Unaffiliated" (but not their parent categories, to avoid double-counting); for Mormons I did not distinguish between LDS and the other 4% of the total Mormon population (which must largely be made up of Community of Christ Mormons*); I broke the Orthodox categories into non-Greek Orthodox and Greek Orthodox. That sounds jumbled, but take a quick look at the categories here and it'll make sense.

Then I compared the predicted and actual percentages of 3+ kid households by subtracting the predicted from the actual. So if, based on level of affluence, 10% of the Catholic population is 'expected' to have 3 or more children at home, but in actuality 15% of the population does, the Catholic population has a +5 propagation 'score'. They do more procreating than would be expected for how much money they make. The propagation scores:

AffiliationPropagation score
1. Mormon16.1
2. Muslim5.0
3. Catholic3.6
4. Jewish (Reform)2.0
5. Religious (Unaffiliated)1.4
6. Other Christian0.9
7. Greek Orthodox0.6
8. Protestant (Evangelical)0.5
9. Jehovah's Witness(0.1)
10. Jewish (Conservative)(0.3)
11. Protestant (Mainline)(0.9)
12. Protestant (Hist. Black)(1.2)
13. Hindu(1.4)
14. Atheist(2.3)
15. Non-Greek Orthodox(2.7)
16. Agnostic(3.4)
17. Secular (Unaffiliated)(3.8)
18. Buddhist(4.1)
19. Unitarian/other liberal(4.9)
20. New Age(5.1)

Asked in a mock debate awhile back why Christian conservatives should vote for the Mormon Mitt Romney, Jack Cashill, playing the former governor, cleverly launched into how all-American (that is, GOP stalwarts) Mormons are: They're socially conservative, baby-making married members of the middle class and they vote overwhelmingly Republican, so why worry about their theology when you're getting those kinds of results?

Pew didn't inquire about actual income figures, just ranges, so to say that Mormons are the most fruitful when income is controlled for is to make a precise assertion that isn't justified by the numbers available. Also, the use of a linear equation means the range of expected percentages are narrower than the range of actual percentages are. Thus, Mormons, a fourth of whom under the age of 65 have three or more kids living at home, are moderate outliers and so stand out with the highest score by a long shot.

Further, I removed the 65+ population of each group, reasonably assuming that the vast majority of them do not have children living at home. But I couldn't make an age adjustment for income without more complete data on all of the individuals surveyed. Since people between the ages of 18-64 have considerably higher average incomes than people 65 and older do, the relatively elderly groups--Reform Jews, Conservative Jews, and Greek Orthodox, specifically--actually appear less wealthy than they actually are (even though they are the highest, third highest, and fourth highest groups of the twenty, respectively, in terms of income!) and so are 'expected' to have more children than would likely be the case if income was adjusted to account for age.

That said, I suspect that if income and age were controlled for, the fecundity rankings (the scores, again, are the product of a linear regression based on data entirely from the categories represented--the rank order is more meaningful) would pretty closely resemble the table presented above.
This suggests something that comes as little surprise--traditional religious groups do a better job spreading their genes than religiously progressive and irreligious groups (I'm including most American Buddhists here, over half of whom are white) do, even when income is roughly adjusted for.

Some other points of interest:

- The Hindu and Jewish income distributions are very similar. Together, they comprise an affluent 2% of the US population!

- Buddhists are less likely than members of the general public to make under $50,000 a year (44% to 53%) and more likely than members of the general public to make over $75,000 a year (39% to 31%). I hear Dave Matthews mixing it up a little, singing "What I want is what I've not got, but what I have is what I say I don't want!"

- Subjected to a fairly religious upbringing, in common parlance I've often heard "atheists and agnostics" grouped together as a single entity representing the non-believer, as though they're the same thing. The 'opposition' usually looks less diverse than the 'coalition' does, whether its religious beliefs or any other arena in which groups compete. I've been surprised on more than one occasion during the election season by politically active friends on the left (as most of mine are) who saw the Republican field as largely indistinguishable (with the exception of McCain and Paul to some extent). This surprise, in spite of how, with the exception of Biden and Kucinich, the Democratic crop seemed like a gaggle of doppelgangers to me!

Well, A&As are strikingly similar with regards to the major demographic markers: In terms of income levels, educational attainment, race (mostly white, although Asians are overrepresented), age, geographical distribution, marital status, and the number of children at home, gender (eep, don't say anything about the pious being pinheads now, as there are twice as many male A&As as there are female A&As!), etc. That is, on every single attribute Pew examined, atheists and agnostics are virtually identical in their distributions (with the minor qualifications that agnostics are a bit more likely to be boomers than atheists are, 22%-16%, and are slightly less likely to be male than atheists are, 64%-70%).

- The Reformed Jewish population has more children across the board than the Conservative Jewish population does, with the exception of 4+ children families--a category that represents only 1% of the Reformed and 2% of the Conservative populations. That certainly surprises me. Is it just orthodox Hasidic Jews, then, who tend to have large immediate families?

- Hispanics make up almost one-third of American Catholics, even though they only comprise 12% of the population. They make up single-digit percentages of all the other major religious affiliations with one exception--Jehovah's Witnesses. Twenty-four percent The Watchtower people are Hispanic. I had no idea.

- If you want to marry Manjula but won't worship Vishnu, you're just about out of luck. Excepting A&As, Hindus are the most gender-skewed group, with a 61%-39% male advantage. Hindus have the highest marriage rate of all (79%), as well as the highest rate of marriage to someone who shares their religious affiliation (90% of those who are married). Thus, 71% of Hindus aged 18 or older are married to other Hindus.

For that 71% to be reached, 35%-36% of the total Hindu population must be comprised of Hindu males married to Hindu females and conversely another 35%-36% of the total Hindu population must be comprised of Hindu females married to Hindu males. That means 3%-4% of the total Hindu population consists of Hindu women either married to non-Hindus or still on the market. Slim pickings!

* The CoC is headquartered in the eastern part of the Kansas City metro area. Theologically, it is closer to mainline Protestantism than the Church of LDS is.


Sleep said...

Im surprised that the fertility of Greeks isnt lower. Most Greek families dont seem to have many kids. Once, I heard someone tell me "you're young, for a Greek". Also, in my experience, Orthodox Jews have huge families but also huge turnover rates, with the net result that they're barely breaking even in terms of population growth. Hasidic Jews, on the other hand, are off the charts. (Orthodox and Hasidic are not synonyms.)

Audacious Epigone said...


Greece's TFR is one of the lowest in the world, but as you know, there's a significant 'diaspora' in countries like the US and Australia. Is the story the same among these Greek populations?

I'm thinking My Big Fat Greek Wedding, where the family expects the woman in her thirties to have made babies by now! Removing the 65+ portion of each group, and without any age adjustment, Greek Orthodox are the 9th most likely to have 3+ kids living at home (out of 20). That certainly suggests a TFR closer to that of the US (around 2), than of Greece.

Sleep said...

Sorry. By Greeks, I meant Greeks in America. I have no exposure to the culture of Greece itself.

On the subject of Hasidic Jews, I thought this was interesting:

Al Fin said...

Getting sidetracked on Greeks? Is it all Greek to you?

Seriously, Mormons and Evangelicals along with a thin thread of Catholics (and a few Orthodox Jews and Americanized East Asians) may provide a just slim enough productive class to ease the US into the next era.

What is the next era? It's the accelerated biotech, nanotech, advanced machine/human cognition era that acts as a force multiplier for the human competence that is there.

It's one of the stages leading into what I call the "next level."

Audacious Epigone said...

62% living below the poverty level? Is that correct? That's almost six times that national rate.

Audacious Epigone said...


That's an interesting way of looking at it, as though the productive elements of society are in a race against time, working feverishly on the azoth as the hordes are at the gate, in the courtyard, down the hall... That might sound over-the-top, but it's the sense I get, too.

agnostic said...

Re: Manjula --

Asian-American outmarriage rates

36% of US-born-and-raised Indian females are married to US-born-and-raised Whites. And that's just the marriage rate -- standards tend to be more liberal when you're just dating.

So cinnamon ain't afraid-a a little sugar.

Anonymous said...

The PEW REPORT is "nice" for several reasons, but INACCURATE.

For instance, Hispanics actually constitute between 30% and 35% of U.S. Jehovah's Witnesses, and African-Americans constitute between 20% and 25% of U.S. Jehovah's Witnesses.

Here is a much more accurate analysis:

Anonymous said...

I think your Manjula analysis may need some additional work.

A non-negligible percentage of married Hindu males are married to women in overseas (most likely India), and thus their brides will not be covered by the survey.

Audacious Epigone said...


The Jehovah's link isn't accessible. Have a link without restricted access? This wikipedia article puts the black percentage at 37%, far higher than Pew's 22%. But the study referenced in the wiki article also found 71% of Witnesses to be female. Is the gender distribution really that skewed (Pew found it to be 60%-40%, female)?


The Pew survey suggests about 1.2 million Hindus in the US, to 1.7-2 million people in the US of Indian descent. Most of the Hindus surveyed are foreign-born (86%). Any idea how much of the US-born Indian population is other than Hindu (Muslim or Catholic, like Bobby Jindal)?

The marriage rate to Indian men for all Indian women in the US is the highest (93.6%) of any of the 6 largest Asian groups (both genders) is at least partially a result of how the second generation of Indian-Americans are coming into marriage now, and as you point out, they are grinding plenty of salt.

So the Manjula reference still works pretty well, since she was raised in India, even though I didn't think it through with that in mind.


Good point. I made note of it in the body of the post. The survey doesn't report on the status of spouses. Any idea on how big the slice of Hindu men living in the US with spouses back home is?

Justin Halter said...

I postulate that the Asian over-representation among atheists is because of the officially atheist stance that the Communist governments of China and Vietnam have taken. Thus, many Asian immigrants simply say about themselves, "I don't have a religion."

Japanese tend to be the same way. While practicing the cultural Shinto back home, in America they "don't have a religion" because to them, only people that belong to a Christian or Buddhist sect have a religion.

dave in boca said...

I married a Greek Orthodox girl and we have one child. I come from an Irish Catholic background and the average for an Irish Catholic in Winnetka where my daughter attended Faith, Hope & Charity in the nineties was about five kids per family, even though the IC types there were rich & owned the Cubs, the Bears, the Trib, and various other Chicago institutions. Their intermarriage rate was ferocious and my Niki had one class in which thirty [30] of the kids, all of rich Irish parentage, were at least second cousins [most were first cousins & a few twins]. I used to joke about "one blue eye & one green eye." The owner of Dixie Cup had twelve kids and counting---sort of like Mormons in heat!!!

Not to put too fine a point on it, but a lot of Mormons are of Irish or Scottish descent, with Steve Young [QB of the Forty-Niners & ggg/son of Brigham] sort of typical Mormon Irish.

:ast night, Fox's "Red-Eye" had a handsome Mormon dad named Troy Dunn who had seven kids. He looked like a refugee from the potato table.

Audacious Epigone said...


I was thinking the same thing regarding Taoism or Neo-Confucianism (that's religiously dead, I know) as well. In addition to Shintoism/animism (and probably Buddhist practices for many Asians), these are not really religions in the same way that the deistic major Western religions (and Hinduism) are. So maybe white Americans see themselves as religious when they adopt Buddhist practices, but Asians who are as or more 'pious' in the US say they have no religion.


Right, that Irish/Scot Mormon descent means Mormons are ethnically sort of 'out of place' in the Southwest, and considering only residents of European descent.

Is the rate of Catholic bloom continuing into subsequent generations, or have those families you grew up around moved more in the one-two child direction?

Dr. Michael Blume said...

Thanks for the interesting post and thoughts. In some European countries like Switzerland and Austria, we have censusses which include religious affiliation. Therefore, we are making progress in tracking the "reproductive benefit of religiosity", which is important not only to analyze our present (like the higher birth rates in the religious dynamic US compared to European countries), but also to understand the evolutionary past of religiosity.