Notes From The Bell Curve

(Written 5/09/18)
I've been working on this book since I got it for Christmas and I've finally finished it. Here are notes I wrote down with each chapter, all in one place.

Chapter 1:
Cognitive classes are getting increasingly segregated, as of 1990, the odds of half of 12 random Americans having a college degree was six in a thousand, yet for most college graduates that's all or almost all their closest associates.

Chapter 2:
A hundred years ago high IQ folks were largely scattered throughout society. Now, high IQ folks are very concentrated in high IQ professions.

Chapter 3:
Intelligence (g) is the single best predictor of job performance for almost every job, it beats every other known variable. It beats an interview, reference checks, college transcript and even beats tests designed specifically for the skills that job requires. Hiring people based on IQ tests would add billions to the economy each year, but it is currently illegal.

Chapter 4:
Intelligence has become increasingly important to employers. IQ is currently estimated to be 60% determined by genes. (One way to estimate this is studying identical twins separated at birth.) High IQ people are becoming increasingly segregated from average/low IQ people. High IQ people make way more money and high IQ professions are the most lucrative. A lot of the increased income correlated with more years in education could be because years of education is correlated with IQ scores, smarter people stay in school longer. Super smart people are getting richer even as income for everyone else has stagnated.

Chapter 5:
IQ is generally set for life by age 10. The following statistics are for non-Latino whites only: Poverty is strongly associated with childhood socioeconomic status (SES) but even more strongly associated with intelligence. Using regression analysis, IQ is much more of a cause of poverty than childhood SES. IQ strongly predicts poverty among high school (no more no less) graduates. Single mothers have ridiculous poverty rates and the rise in single motherhood is largely responsible for the increase in childhood poverty rates. Even with single mothers, IQ was strongly predictive of poverty, with poverty rates ranging from 70% to 11% at each end (+- 2 standard deviations) of the IQ spectrum. Until the 1960s it was completely acceptable to distinguish between the "deserving" and the "undeserving" poor. Now saying someone is poor because of their actions is "blaming the victim."

Chapter 6:
As high school graduation has become the norm, dropouts have become much more self-selected for low IQs. IQ is way more important than SES in predicting high school dropout rates. Less than 1% of dropouts meet the smart but disadvantaged stereotype (top 25% IQ, bottom 25% SES). Flipping those around, only 12% of "dumb rich kids" end up getting college degrees. IQ is more important than SES in predicting who gets college degrees.

Chapter 7:
Low IQ is associated with unemployment (20% vs 5% chance of being out of labor force for a month (white men) for +-2 SDs). Taking IQ into account, parental SES is actually slightly inversely correlated with being out of the labor force. Dumb people are more accident prone at work. Smart people get in less wrecks.

Chapter 8:
(Statistics for Whites only still.) Marriage rates are the one thing where it appears education and not IQ causes the correlations. High IQ folks stay in school longer, pushing back marriage. Among only HS grads, IQ was actually positively correlated with marriage. Dumb people get divorced more. IQ has a large effect on illegitimacy rates, over and above the mother's childhood SES, ranging from 34% to 4%, controlling for SES +- 2 SDs. Flipping that around and controlling for IQ and letting SES vary, it ranges from 19% to 8%, still significant but not nearly as much so. Growing up without a father (but not so for mothers) more than doubles chance of having an illegitimate baby, even controlling for both IQ and SES.

Chapter 9:
College educated whites basically never go on welfare. The number of families on welfare tripled in the late sixties-early seventies. Both IQ and parental SES have significant independent effects on whether a women will go on welfare, with parental SES: 40% to 5%, and IQ: 28% to 8%, SES actually has the larger effect.

Chapter 10:
IQ has a significant effect on rates of underweight babies. Controlling for IQ, younger mothers' babies weigh just as much as older mothers'. Most malparenting situations come from those in the lowest social classes. Mother's IQ is a stronger predictor than mother's SES background of the worst home environments. In public dialogue, problems experienced by poor children are almost always attributed to the poverty itself, but many of them may really be not a result of the poverty itself but today's correlates of poverty (e.g. living without a father). In predicting child's IQ, the mother's IQ was more important than her SES background: 30% to 2% for IQ vs 10% to 4% for SES (+-2 SDs, holding the other variable constant in each of course) chance of having a child in the bottom 10% in IQ.

Chapter 11:
IQ is a better predictor than SES at predicting criminality. People with criminal backgrounds often drop out of school but dropping does not appear to increase criminal behavior, they're already exhibiting criminal activity before they drop out.

Chapter 12: The direct relationship between SES and voting is overblown. It is predominately education, not income, that links voting and SES. A poor but well-educated shop clerk is more likely to vote than a less educated, rich businessman. Since years of education is largely caused by IQ, it is likely that IQ is the real link between SES and voting. Also, knowledge about politics is predicted by IQ and interest in politics is predicted by IQ, so basically just IQ, and knowledge about politics is obviously correlated with voting.

Chapter 13:
Average IQ: whites: 100, asians: 103-106, blacks, 85-88. The difference in IQs is not due to cultural bias, blacks do as well or better as their white counterparts on questions with cultural content. Blacks do about as well in the workforce as their IQ would predict, which is more evidence the difference is real. It's not a difference in effort because there's a difference in the backwards digit test but not the forward (the backwards is twice as g loaded). Additionally, whites have faster reaction time but slower movement time and reaction time is much more correlated with g than movement time. SES explains 30-40% of the difference in IQ (probably less since SES is largely a result of parental IQ). Black/white IQ difference is consistent over all SES levels (90 vs 105 a the 50th percentile SES 100 vs 115 at the 90th). For this to be purely environmental the average black environment must be at the 6th percentile of whites. For children adopted by two whites: if both biological parents were black: 89 IQ, 1 black 1 white, 99 IQ, both white biological parents, 106 IQ. Adopting parents were all upper-middle class.

Chapter 14:
Average IQ of college graduates is about 114. After controlling for IQ, whites and hispanics have the same chance of getting a college degree, blacks have much more. Blacks are very overrepresented in high-status occupations (10% for whites vs 26% for blacks, for those with the average IQ of those in high status jobs (117)). 91% of the annual wage difference between whites and blacks disappears after controlling for IQ. However controlling for education does not erase near as much, blacks make less than similarly educated whites. This is a result of affirmative action, black college graduates have lower average IQs, which is the real driving factor behind wages. Stats: blacks on average make 77% as much as whites, controlling for education, they make 80%, controlling for IQ, they make 98%, controlling for IQ + gender they make 101% as much. Controlling for IQ on family income and poverty rates shrinks the gap substantially as well, but not as much as for individual income. (poverty rates: white: 7%, black: 26%, 100 IQ white: 6%, 100 IQ black, 11%.) With IQ 100, blacks have a 15% chance of being unemployed vs 11% for whites. Makes you think lingering racism, but then there's no explanation why it didn't affect individual income from earlier. The huge thing: controlling for IQ erases almost none of the difference in marriage rates. (100 IQ age 30: whites: 79%, blacks: 58% (1991)). Historically, the black-white difference in marriage rates was small until the early 1960s and then widened. Controlling for IQ doesn't do much for differences in illegitimacy rates either (100 IQ black women: 51%, white woman: 10% chance of bearing children out of wedlock). Black mothers are more likely to be on welfare after controlling for IQ (IQ 100: 74% vs 56%). These profound cultural differences in relation to marriage probably explain the leftover differences in family income and unemployment rates after controlling for IQ. Most of the difference in delinquency rates is explained by IQ independent of SES.

Chapter 15:
America has likely undergone dysgenic pressures since 1900. Differential fertility: women with more education have less children. Woman HS dropouts have 2.5 kids while BA's have 1.5. Since education is largely a result of IQ, dumb women are having more kids. Lower IQ women start having their children earlier, which is dysgenic since it results in more generations of lower IQ people. A large welfare state causes immigration to be dysgenic, as low-ability workers come from countries with less extensive welfare states. America has changed from the land of opportunity to the land of the safety net. White women have 1.89 children, blacks have 2.23. Controlling for the mother's education removes most differences in white-black fertility rates.

Chapter 16:
82% of people below the poverty line have below average IQs.
88% of women chronically on welfare had below average IQs.
94% of permanent HS dropouts had below average IQs.
85% of illegitimate children are born to below average IQ mothers.
93% of children who lived in poverty their first 3 years of life had below average IQ mothers.
94% of the dumbest decile of children had mothers with below average IQs. Of the small percentage (2.2%) of men who didn't work without a reason (like being in college or an injury), the vast majority had low IQs. 50% were in the bottom decile of IQ and 84% were below average in IQ. Yet, overall, low IQ men still work full time almost as much as everyone else and the majority of low IQ people are employed, out of poverty, not on welfare, and were married when they had babies (as of that year at least). Controlling for SES does not remove the influence of IQ on most all of these variables, on the contrary, controlling for IQ often gets rid of the effect of SES.

Chapter 17:
Fixing clear nutritional deficiencies can increase IQ by a few points (not clear how much exactly, maybe 6ish). America's school system has advanced enough that variations between the best and worst public schools have no explanatory power over differences in students' IQs. By age 5 or 6 there is little you can do to increase IQ. Intensive schooling will not cause lasting increases in IQ. Programs like Head Start and other even more intensive programs aimed at poor toddlers have produced basically no lasting IQ gains, they always start to fade away after a couple years. Adoption at birth from a bad environment to a good one can have real lasting benefits for IQ (like 6 points). Environment does influence IQ (roughly 40% of its variance is explained by environment) but we haven't figured out how to really improve it short of adoption at birth. Intensive schooling programs don't seem to do much to overcome a poor home environment.

Chapter 18:
While SAT scores were plummeting from the 60's to 1980. NAEP scores remained basically flat (NAEP is a nationally representative sample while SAT test takers were and are disproportionately cognitively gifted). SAT scores mostly went down on the verbal portion. The education reforms of the 1960s, motivated by egalitarianism, dumbed down the curriculum. Lower IQ students were mostly unaffected by this change, but it had the affect of possibly not fully realizing the brighest students' potentials. In 1993, 0.1% of the money from the ESEA went to programs for the gifted, 92.2% went to programs for the poor. Most (almost all) federal education dollars are awarded solely based on need and thus generally go to low IQ people. American culture prefers egalitarianism and pumping up people's self-esteem independent of actual value to prizing excellence. Developing the intellectual elite is important to all of society, not because they "deserve it" (since being smart is largely a gift), but because the future of society will largely depend on them. There is only so much low IQ people can do, even with all the help of the federal government.

Chapter 19:
Blacks get what amounts to a bonus of 180 points on their SAT scores in college admissions. Blacks make up 12% of the average student body but are 52% of the bottom 10% IQ of the student body. Blacks' higher college dropout rates mostly go away when you control for IQ. A black person's college degree is devalued in the corporate marketplace now. A low SES white still had to be 0.58 standard deviations above a high SES black to have the same acceptance rate. Black college grads are generally a full SD below white grads.

Chapter 20:
White elites like affirmative action while working-class whites do not. General cognitive ability tests are banned to use in the hiring process. Basically none of the legal (read: job-specific) tests are more predictive of job performance than a general IQ test. Most of their predictive power comes from whatever extent they capture g. Controlling for IQ, underrepresentation of blacks in professional/technical jobs was gone by 1964, prior to the passage of the Civil Rights Act. Affirmative action hires have significantly lower job performance. Our laws assume that all groups should be equally good at everything, and if they're not, it's prima facie evidence of discrimination, even if there is no actual evidence of any. Equality of opportunity is much more reasonable a goal to strive for, with much less deleterious effects.

Chapter 21: No notes.

Chapter 22:
Egalitarianism underestimates the important of the differences that separate human beings. Neighborhoods have been stripped of many of their traditional functions, which used to give many people a sense (and reality) of value. The criminal justice system needs to be made simpler, in a way that is more of a disincentive to low IQ folks with short time horizons. There needs to be incentives/rewards for marriage (single mothers should not get child support). America's policies actively encourage dysgenic breeding.

Afterward:
Public policy sees society as the cause of all problems, government as the solution, and the absolute malleability of people is taken for granted. One of the earliest findings about mental tests was that the results of different tests of apparently different mental skills are positively correlated. The relationship between IQ and all the social behaviors discussed above is much more powerful than the relationship linking those behaviors with the usual suspects (education, social status, affluency, ethnicity). The presence of the biological father in the home has many important positive effects on children independent of SES.