Less variation in family size

Using General Social Survey (GSS) data, I calculated the mean number of offspring for women ages 45 to 64 for each decade since the 1970s. I display these numbers below, along with standard deviations--a measure of variation--and estimates of kurtosis--the degree to which a distribution is flatter or more peaked than a normal distribution:

1970s
Mean 2.77
SD 2.04
Kurtosis .19

1980s
Mean 3.07
SD 2.05
Kurtosis -.09

1990s
Mean 2.54
SD 1.75
Kurtosis .86

2000s
Mean 2.14
SD 1.54
Kurtosis 1.38

2010
Mean 2.05
SD 1.60
Kurtosis 2.35

We all know that family size has shrunk over the past few decades: according to GSS data, from 2.77 children in the 1970s to 2.05 last year. The standard deviations indicate a reduced amount of variation in completed number of offsapring. In the 1970s, SD was 2.04, meaning that if we grabbed two random  women who had completed their families, our best guess is that one mom would have two more kids than the other. Moving forward to the 2000s, SD has dropped to roughly 1.5 which tells us that the two hypothetical moms differ by one and a half kids. In other words, families have become more similar in size. They are more and more converging on the number two.

If the kurtosis number is one or greater, that means that the distribution is more peaked than a normal curve. While there is no problem through the 1990s, one appears in the last decade. What this means in plain English is that a lot of women are having two children, and that puts a skyscraper right in the middle of the bell-shaped curve. In the 1970s, 24 percent of women had two children. By 2010, it was 34 percent.

Why am I interested in this? Well for one thing, reduced variation in family size means that people are contributing a more equal amount of genes to the next generation than in the past. A few decades ago, some people would have zero kids, some would have ten. Of course, we still have diversity, but there is greater convergence on having two offspring. If that convergence became complete (it won't) every woman would have two children and would contribute the same number of genes to the next generation. Since almost all children (not including fetuses) nowadays make it to adulthood (thank God), there is even less differential mortality than differential fertility. It looks like the evolutionary process ain't what it used to be.     

But what about the male contribution, you ask. That's next.

Racial/ethnic differences in homosexuality

A reader suggests that homosexuality might be more accepted in the Hispanic community because of a higher prevalence. While working as a teacher, he noticed quite a few effeminate Latino boys.

The GSS asks repondents about the gender of their sexual partners. I calculated for each major racial/ethnic group the percent who are gay or bisexual (i.e., the percent whose sexual partners are exclusively same-sex or who are of both sexes). I combined these two categories to boost sample size.

Percent gay or bisexual

Males (sample size = 7,753)
Whites 3.4
Blacks 4.4
Mex-Ams 4.5

Females (sample size = 8,694)
Whites 2.7
Blacks 3.0
Mex-Ams 1.8

Mex-Ams do have an apparently higher level of male homosexuality, but there are no statistically significant differences among any groups at the 95 percent confidence level. (The sample sizes for minority homosexuals are small.)

The black/Mex-Am female difference is statistically significant at the 90 percent confidence level. Lesbianism seems to rise with the independence/dominance of the females of a racial/ethnic group.

Mexican-American IQ

In a previous post, I used GSS vocabulary data to show a 10 point improvement in IQ among Mexican-Americans over the past four decades.  An increase from 85 and 95 is striking, but I didn't take the time to see if the change is statistically significant.

Setting the white IQ at 100, here are the means by decade for people born in America who are of Mexican descent. Sample sizes are shown in parentheses:

Mean IQ

Seventies 85.4 (12)
Eighties 85.6 (106)
Nineties 91.7* (140)
Two-Thousands 94.1* (182)

*significantly higher than the 80s mean

The means for the Nineties and the Two-Thousands are significantly higher than the Eighties' mean but do not differ significantly from each other. The estimate for the past decade is a bit lower than in my last analysis; this is due to the addition of 2008 data. (The mean for 2010 is 91.6, n = 46.)

A mean in the low nineties is not inconsistent with published studies. The low mean of 85 observed in the 1970s and 80s might be due to a lower average level of education. The average respondent in a GSS survey is in his mid-40s which means that if he participated in a survey in the 70s or 80s, he would have gone to school in the 1930s and 40s.  

Hispanics are natural social... liberals

Some Republicans claim that immigration from the south is a good thing since Hispanics are natural social conservatives. Just the other day Rick Santorum was arguing on Fox that emphasizing social conservatism is an essential way to attract Latino voters.

Such a view is just plain idiotic. I regularly check attitude surveys and cannot remember the last time I saw Hispanics with more socially conservative views than whites. Here is the latest from Pew:

The share of Hispanics who think homosexuality should be accepted is six points higher than the white number. Latinos line-up closely with self-described Democrats. They are almost 30 points away from conservative Republicans.

You might have expected their Catholicism to shift them right. But look at the Catholic estimate: 64 percent feel that homosexuality should be accepted. There is a huge gap between Catholic doctrine and social reality.

Liberals are not smart, but they are geniuses compared to Republican leaders. They know that, chances are, each new Hispanic citizen is a new Democrat. And that ain't gonna change.   

Hispanic women and waist-to-hip ratio

Using MIDUS data, I regressed waist-to-hip ratio (WHR) onto race/ethnicity and body mass index (BMI). (I figured fatter people have higher WHRs.) Here are the results for 1,974 women:

Standardized OLS regression coefficients

Hispanic .06*
Black -.02
Asian .01
BMI .32*

*statistically significant relationship with WHR

Fatter women do indeed have fatter waists compared to hips. The three minority groups are all being compared with non-Hispanic whites. Latino women have higher WHRs than their white counterparts. The other two racial groups do not differ from whites.

Consistent with my casual observations, Hispanic women are less curvy. Waist size is more similar to hip size. The curvaceous woman is more of a white thing.

Prenatal testosterone and sexual jealousy

In a new study of 280 people, researchers found that a measure of feminized brain structure (i.e., 2D:4D digit ratio) was associated with less distress at the hypothetical scenario of sexual as opposed to emotional infidelity by an intimate partner. The relationship was observed for men and women together, but not for each group individually. This relationship for the combined groups disappeared when gender was controlled. This supports the conclusion that males's higher level of prenatal testosterone masculinizes their brains making them more focused than females on sexual infidelity.  Men are thought to be more sensitive than women to cues of sexual infidelity to prevent cuckoldry, while females are more sensitive than males to cues of emotional infidelity to prevent loss of parental investment.

The evidence keeps pouring out that gender is not a social construction (as if any reasonable person would think it is).

He is that which is not contingent

I'm traveling and have not had opportunities to post. I wanted to do an amusing post on racial differences in attitudes toward which race is sexually most well-endowed, but technical problems are delaying that a bit.

Allow me instead to talk about something that requires no data and which is far removed from sexual endowment.

In response to the claim that God explains the existence of the universe, atheists like to ask "But who made God?" This question misses the point. The universe and everything in it is contingent: It exists, but it could not exist. In fact, it seems more likely that nothing would exist. Nothing is simpler than something. Anything that is contingent requires a cause for its existing rather than not existing. The contingent universe requires a necessary entity to explain it. By necessary, we mean that which is not contingent; that which does not rely on something else for its existence. We call the entity which causes the contingent universe to be, God. He was not created and could not possibly be created. To ask where He came from is to not understand what He is by definition.  

Thinking that whites are smarter than blacks

The Detroit Area Study asked a sample of 384 blacks and whites whether the word "intelligent" better described blacks or whites. Respondents could also answer "equally descriptive" or "not descriptive of either."

Here are the responses in percentages:

Whites
More descriptive of blacks 0.6
Equally descriptive 54.5
More descriptive of whites 33.4
Descriptive of neither 11.4

Blacks
More descriptive of blacks 19.7
Equally descriptive 68.4
More descriptive of whites 11.8
Descriptive of neither 0.0

It is not surprising that "equally descriptive" is the most popular answer given by both races. Not only is America's culture egalitarian, there is also a tendency to interact with people similar to yourself.  Smart people often find themselves among smart people, and slower folks with people like themselves. Until I lived in Newark, I had only met middle-class blacks, and was stunned by the "differentness" of inner-city African Americans. Whites with little exposure to poor blacks might sincerely believe that IQ averages are the same.

Blacks spend more time with whites than the reverse, so they are more familiar with the other race, which might help explain that there are actually a few (11.8%) who admit that whites are more intelligent.

Of course, it's taboo to publicly claim that whites are smarter, but one-third of whites manage to say this to the interviewer.

Images of God and crime

The sociologist Rodney Stark claims that a religion's image of God has an impact on the conduct of adherents. People will watch their behavior more closely if they see God as a ruler rather than a friend.

GSS respondents were asked how likely it is that 1) a friend, 2) a master, or 3) a king come to mind when thinking of God. Answers ranged from very likely to not likely at all.

I calculated the contigency coefficients and p-values for the relationships between the answers to each of the three questions and whether you have ever been arrested for a crime. The sample is 388 white people. The relationship between arrest and seeing God as a friend is not statistically significant, but it is for both seeing God as a master and as a king. People with this type of image are less likely to have ever been arrested. On the other hand, the strength of the connection is weak. The coefficient (which is like a Pearson correlation) is .12 for master and .17 for king.

In sum, imagining God to be a powerful authority predicts less crime. It is possible that the trend among Christians to see God as a buddy works against improving behavior.  

Jared Taylor on transcending race

Jared Taylor has a new book on white identity. I look forward to reading it.

In the highlights of his ad, he claims that only whites have a desire to transcend race. The Detroit Area Study asked people (sample size = 378): "Generally speaking, do you favor full racial separation, integration in some areas of life, or full integration of the races?"

Sixty-five percent of whites want full integration. So Taylor seems to be right about whites. But what a about blacks? Seventy-two percent want full integration. That doesn't sound like "whites are the only race..." People don't always act as if they want to transcend race, but Taylor also makes claims about aspirations.

Marital decline among women

A reader asked if I could post the graph for marriage trends for women in their 20s. You can see that the percent married has dropped dramatically over the past 40 years. The norm was reversed in a historically short period of time. The rate is higher than for men of the same age, but women experienced a sharp drop in just the last two years: The percent married fell from 26.0 percent in 2008 to 22.5 percent in 2010.

The decline in marriage continues

This graph shows the decline in marriage among men in their 20s. Fifty-nine percent in 1972 were married. That has plunged all the way to 17 percent in 2010. The drop between the two most recent surveys--2008 and 2010--was 2 points.

Fifty-two percent of men in their twenties who are married attend church more than once a month. In the past year, my religious 21 and 22 year old nephews (the latter is on his way to med school) got married to great girls of the same ages. God bless 'em.

UPDATE: Audacious Epigone has reminded me that there are bloggers who doubt the validity of GSS data. Why haven't I seen them expressing their doubts about the data in the comment sections of this post?

Job prestige for teachers

The multi-decade refrain that teachers don't get enough respect is nonsense.

Sociological research has documented that job prestige is largely a function of the level of education and income associated with the particular occupation.

I regressed GSS job prestige on respondent's education and income (sample = 27,972). Next, I calculated the mean income and educational level for American secondary school teachers. Finally, I plugged the means into the equation which yielded an estimated prestige score of 51.9. So based on a model of job prestige, secondary school teachers are expected to have a score of around 52. What score do they actually have? It's 66. That means that people assign greater prestige to teaching than what would be predicted from a model based on educational level and income. Fourteen more points--that is a lot. (As an example, teachers are held in the same esteem as professional athletes (score 65) but they should be looked at like social workers (52)). 

Teachers get much more respect than other people with jobs requiring similar levels of education and producing similar levels of income. The people working these jobs, not teachers, are the ones getting screwed.

Housework and sex frequency

Roissy has claimed that husbands who do more chores at home have sex less often because it is a turn-off to their wives.

The GSS asked respondents how often they have sex, and how many hours of houeswork they do in a typical week. Sex frequency is influenced a great deal by age, so I include it as a control:

Standardized OLS regression coefficients (sample size = 218, married men)

Hours of housework -.05
Age -.45*

*statistically significant

While age exerts a powerful effect on frequency of sexual intercourse, housework is unrelated. (It might have a small negative effect if the sample size were larger).

Wives are less likely to want sex if they are tired or if they are in a bad mood. They are more likely to be in a bad mood if they feel that their husband is not pulling his load. (I'm skeptical of the view that what a woman finds attractive in her man is his exploitation of her). The chance of her wanting sex will be higher if she gets manly vibes from her husband. So the best plan for a husband is to show his wife that he is doing his fair share, but to focus on manly types of work. That means work that is physical or technical. Much of it will be outside. If you don't know how to fix stuff, like a car--you're smart--learn. Avoid any type of work, except for barbecuing, that requires an apron.

The goal for a good marriage is to be traditional but to be just. In Bruce Charlton's words, be a patriarch.

Decline in confidence that God exists

Once every two years the Inductivist gets giddy with excitement when the new GSS comes out. As a first plunge into the new data, let's look at trends in confidence in the existence of God.

The shrinking purple section (if I were a girl, I'd know the precise name for that color) indicates that "knowing that God exists" has been on the decrease over the past decade. On the other hand, I wouldn't make any long-term predictions. Most of the experts sixty years ago predicted the imminent death of religion.

Predictors of closeness to whites among whites

GSS respondents were asked how close do they feel toward whites. Limiting the analysis whites, I regressed the answers onto a list of predictors. Here are the results (sample size = 2,806):

Standardized OLS coefficients

Female .09*
Age -.03*
South -.02
Size of place .02
Education .02
Income .05*
IQ -.07*
Church Attendance .08*
Jewish .01
Conservatism .00

*statistically significant

Much of this surprises me. Being from the South or a small town is unrelated to liking whites. So is years of education or level of conservatism. Jews like whites better than others at the bivariate level, but the relationship becomes non-significant in the multivariate analysis.

What predicts greater closeness to whites is: being female, being younger, having more income, being less intelligent, and more church attendance. The gender finding replicates other research that has found that women have more race affection than men. Who would have guessed that younger people feel more affection for whites than their elders?

Although all of the predictors are weak, gender, religiosity, and IQ are comparatively the strongest. Why am I not surprised that smart people like whites less? IQ exerts an effect while education does not. Interesting.

Race and ethnicity among Hispanics

As a follow-up to the last post, I calculated how warmly Hispanics of three different races (white, black, other) felt towards generic whites, blacks, and Hispanics (sample sizes = at least 200).

Overall, Hispanics are the best liked group (not surprising); blacks are the least liked. This is the same pattern we saw in the last post with all races.  

White Hispanics like Hispanics better than they like whites. They like their Latinness more than their whiteness. People usually like their whatever-elseness more than being white if they have a choice. You distance yourself from whiteness if you can. Whites also like their own race the least (the difference is very big). I suspect that while they describe their race as white, many answer the closeness-to-whites question thinking of whites as "Anglos." 

Black Hispanics like blacks better than Hispanics. They like their blackness more than their Latinness. Black trumps Latino in the race/ethnic hierarchy.

Other-race Hispanics like whites more than blacks, but the means for both are low. They like Hispanics much more than either blacks or whites. The 2.5 point gap between the mean closeness towards blacks and that towards Hispanics is huge. 

Feelings of warmth among races

This table shows the mean scores of how warmly one group feels toward another, based on GSS answers from 1 (very cold) to 9 (very warm). The columns are the judging groups, while the rows are the target groups. "Whites" are "blacks" are non-Hispanic; "Hispanics" are "other race." "Asians" in the judging group are Japanese and Chinese Americans, but are simply Asians in the target group.

Going along the rows, you can see that Asians and Hispanics tend to be the best liked groups. I suppose this puts a tally mark in the plus column on the immigration question. Perhaps Asians and Hispanics tend to be inoffensive? Blacks are the least liked by everyone (except for blacks of course). Must be because whitey taught the others to hate.

Going down the columns, it is not clear who judges the harshest in general. Asian feelings towards blacks are comparatively cool, but they like Hispanics. Whites judge blacks more warmly than others do--white guilt perhaps--but they are comparatively cool on Hispanics and Asians. I see no evidence here that whites have a unique gift for disliking other races. Unless the race is whites--they have the lowest self-directed mean.

He gets made in our image

This new study supports the view that people make over Jesus in their own image (e.g., blacks believing he was black; Nordics believing he looked Nordic):

In two studies, we examined whether (a) conceptions of Jesus would differ between Koreans and Americans, and whether (b) national differences in self-reported personality and well-being are mediated by the cultural norm for personality and well-being. Because there is only one Jesus, different conceptions held by Koreans and Americans are likely to reflect cultural construction processes. In Study 1, we asked Korean and American participants to engage in a free association task with Jesus as a target. Americans associated Jesus with primarily positive connotations (“awesome”) and rarely with negative connotations (“pain”), whereas Koreans associated Jesus with both positive and negative connotations. In Study 2, we asked Korean and American participants to rate Jesus and themselves using personality and well-being scales. Americans rated both Jesus and themselves as more extraverted, agreeable, conscientious, open, and happier than did Koreans. Most important, national differences in self-reported agreeableness, conscientiousness, openness, and happiness were partially mediated by conceptions of Jesus.

Myself, I imagine him with a hilarious, self-deprecating sense of humor and a nagging fear that he's coming down with something.

Parental income and IQ among teens

Here's a new study of a large sample of teens published in Personality and Individual Differences that shows that the IQs of children from wealthy homes are higher, not because of money, but because of their parents' superior genes.

Parental educational level and family income have been related to individual differences in intelligence. However, large and representative samples are hardly available. Here two samples of young and old adolescents totaling 3233 boys and girls completed an intelligence battery comprising abstract, numerical, verbal, mechanical, and spatial reasoning subtests. Parents’ educational levels, family incomes, and adolescents’ general intelligence (g) were simultaneously related using SEM (structural equation modeling) analyses. The main findings show that (1) parental education strongly predicts family differences in income, (2) family income is not related to adolescents’ intelligence, and (3) parents’ education predicts adolescents’ intelligence regardless of family income. Because it is widely acknowledged that personal intelligence is the best predictor of educational differences, the next causal chain is endorsed: brighter parents reach higher levels of education, which allows approaching better occupations, and, therefore, they can create families with higher incomes. Adolescents from more affluent families tend to be brighter because their parents are brighter, not because they enjoy better family environments.