Saturday, December 05, 2015

Among whites, conservative-liberal fertility difference twice as large as Democrat-Republican difference

Responding to Hail's comment at Steve's, I wrote the following:
Worth reading all of the posts Jayman links to. He has written on this extensively.

It is also worth considering how much the political orientation gap among white Republicans and white Democrats has widened over time. In the mid-seventies, white Republicans were only slightly more likely to self-identify as politically conservative than white Democrats were. That difference has trebled in the last four decades to the point that white Republicans are now far more likely to identify as conservative than white Democrats are.

Political orientation is probably more heritable than partisan affiliation. I'd guess the gap will appear wider on the liberal-moderate-conservative spectrum than on the Democrat-independent-Republican one.

Two variables that are stronger predictors of fertility than political orientation or party affiliation are educational attainment (inversely correlated, especially for women) and religiosity (positively correlated--to the extent that high IQ people who attend religious services regularly outbreed the irreligious at every level of intelligence, social class, race/ethnicity, etc.

Parenthetically, educational attainment looks to be the driving force, not intelligence. Fertility by wordsum score varies little once educational attainment is controlled for, but educational attainment is a strong predictor even after wordsum score is controlled for--put more clearly, educational attainment is 5x as strong a predictor of fertility as IQ is.

The strong inverse relationship between education and fertility shows up strongly on the international level as well.
Instead of guessing, I should do the requisite data delving to evaluate that assertion about political orientation being a stronger predictor of fertility than partisan affiliation for whites. The results, using parameters similar to Hail's*, are as follows (with % of white female population in parentheses):

Extremely liberal (2.5%) -- 1.61
Liberal (12.7%) -- 1.72
Slightly liberal (11.3%) -- 1.58
Moderate (39.9%) -- 2.14
Slightly conservative (15.6%) -- 2.03
Conservative (14.3%) -- 2.34
Extremely conservative (3.7%) -- 2.62

Indeed, the fertility gap appears wider on the liberal-conservative spectrum than it does on the Democrat-Republican one. Hail's results show a white Republican:white Democrat total fertility rate ratio of 2.07:1.76. The white conservative:white liberal total fertility rate ratio is twice as large, at 2.23:1.65.

GSS variables used: POLVIEWS(1-3)(4)(5-7), RACECEN1(1), SEX(2), COHORT(1944-1974), CHILDS, AGE(40-50)

* I used the COHORT variable to pull data from those born between 1944-1974 while being able to exclude survey results from 1994, 1996, and 1998 since in those years "white" includes Hispanic. From 2000 onward, non-Hispanic whites and Hispanics are separated into distinct categories. Consequently, his numbers include some Hispanics, while mine are only comprised of non-Hispanic whites. Also, what Hail terms "political orientation" I regularly describe as "partisan affiliation". I use the phrase "political orientation" to refer to the liberal-moderate-conservative spectrum rather than the Democrat-independent-Republican one.

12 comments:

TangoMan said...

The white conservative:white liberal total fertility rate ratio is twice as large, at 2.23:1.65.

How do liberal Republicans compare to conservative Democrats?

Tangentially, with respect to education attainment, this very much looks like a proxy variable for other factors, factors like more career opportunity, more social mobility, etc which allow women more choices in life, thus lessening the fertility.

Do you believe that GSS has variables which can allow a test of the influence of education versus the influence of other factors on the question of fertility?

Jokah Macpherson said...

It looks like you are still .62 below average, although I can't really cast stones.

Audacious Epigone said...

Tangoman,

I've looked at a few times in the past (here and here). The two most predictive variables in the GSS I've come across are educational attainment for women (inversely correlated with fertility; much less the case with men) and religiosity (positively for both men and women). Wordsum (IQ proxy) is much less predictive than either of these variables are.

Jokah,

My wife is only 25. Her ultimate TFR is yet to be determined, but the smart money says it's yet to be finalized!

TangoMan said...

Can you check to see how liberal Republicans compare to conservative Democrats? How much are conservative Democrats propping up liberal Democrats and how much depressive effect arises from liberal Republicans being mixed with conservative Republicans?

To the educational attainment variable, what are you thoughts of the causality? I suspect that education by itself isn't what is driving lower fertility, rather it's the world that is opened up via education, more career options and perhaps even increased debt load, or delayed marriage, that is lowering fertility. Does GSS have a variable for years of education, so we can analyze whether it is the credential or the education itself which has influence? If there is no change in fertility between 13 years, 14 years, 15 years of education, but then after 16 years, after earning a bachelor's degree, we see a change, then that strongly suggests that it is the credential and not the education which is the operative variable.

Anonymous said...

Someone recently commented on Steve Sailer's blog about the "Janissary effect". It doesn't matter if your side has children if they're converted away..

Anonymous said...

Is this TFR rate gap growing in younger generations? My [hopeful?] impression from looking at my age cohort, "millenials", growing up over the last 10 years has been that the extremely liberal aren't having hardly any children. Perhaps confirmation bias.

JayMan said...

@Anonymous 12/6/15, 1:50 PM:

"It doesn't matter if your side has children if they're converted away.."

As I said over there, don't worry about it. The "gene-environment" correlation on political attitudes is negative (a nonsensical result, which means the true value is zero), as found by Hatemi et al's massive extended twin study.

Environment like school doesn't make one more liberal or conservative.

Audacious Epigone said...

TangoMan,

Yes it does have a variable for years of education.

The credential break is significant, good instinct.

For white women born 1944-1974 who were aged 40-50 over survey years 2000-2014, mean children by years of education:

13: 2.21
14: 2.17
15: 2.22

These are "some college" but no BA.

16: 1.76
17: 1.61

Those are BA.

18: 1.98
19: 1.60

Master's (6 year master's for white women is probably heavily in education, and teaching is a middle class profession that makes it fairly easy for a working woman to juggle while also raising children) doesn't look much different.

20+: 1.15

Doctorate is a barren zone though.

Also, mean number of children for liberal white Republican women (same parameters as above): 1.69

For conservative white Democrat women: 1.65

Audacious Epigone said...

Is this TFR rate gap growing in younger generations? My [hopeful?] impression from looking at my age cohort, "millenials", growing up over the last 10 years has been that the extremely liberal aren't having hardly any children.

Although only 24% of precinct results have come in, there is no reason to think the gap is closing. It may well be opening up further. Mean number of children for liberal white women born in 1982 and after who were aged at least 18 at the time of the survey: .38

For conservative white women with parameters that are otherwise the same: .53

TangoMan said...

Thanks for running that to ground.

So it's not really education attainment which is at work, the expanded intellectual horizons of women are not leading them to reject or delay motherhood, it's some combination of materialism, putting education achievement above marriage, putting career above marriage, waiting around for a better guy, etc all happening while their fertility clock is ticking, which is most pronounced for the doctorates - long years of study during peak fertility.

So what happens to fertility if they marry a good income husband while still young (so there is time for increased fertility to play out)? Kind of related, when Alberta was going through its oil sands boom and there were severe worker shortages, wages increased and this resulted in women dropping out of the workforce, which further restricted labor supply, driving up wages even more, leading to more women dropping out. No other Canadian province saw this happening. Apparently women were choosing to focus on home life when their husbands earned enough to support the family at a middle class level. I'd venture that the same phenomenon played out in the ND during the Bakken boom but their labor shortages weren't as severe.

Women in the workforce, just like immigrants in the workforce, expand labor supply and lower returns to labor. As more women enter the workforce, their wage depression effects impact on everyone else and so force other women to enter because their husband can't earn enough, thus creating a positive feedback loop. This spills over into delayed marriage and reduced fertility.

Does GSS have a "husband's income" variable? Find young women married to husbands who can support them on one salary and compare to other married young women who are working.

Oh well, just rambling at this point. Thanks for that detective work.

Santoculto (santodaemon) said...

Would be interesting analyse singly the fertility rates of HETEROSSEXUAL liberal men and women because higher rates of homossexuality among liberals should decrease their numbers.

Audacious Epigone said...

Tangoman,

There is a workaround in the GSS--look at family income of married women who don't work and you have a good homemaker proxy.

It is worth recalling that the GSS shows clearly that educational attainment is a much stronger predictor of female fertility than it is of male fertility.

Santoculto,

Liberals are nearly 5x as likely to be gay/bi as conservatives are. That's worthy of posting, thanks.