IQ and intelligence
Thread by @PaoloShirasi on Thread Reader App – Thread Reader App Paolo Shirasi Aug 15 • 22 tweets • 10 min read [THREAD]
A VERY BASIC INTRODUCTION TO THE HEREDITARIAN POSITION ON RACE AND IQ.
PART 21: Rebutting Claims That IQ Tests Don’t Measure Intelligence (Or Anything Valuable)
One type of rebuttal sometimes appearing in the popular media, but almost never made by scientists, can be expressed as “intelligence cannot be measured,” and also, relatedly, “IQ tests don’t measure anything but one’s ability to take IQ tests.” These arguments have been scientifically discredited.
It’s been the official position of the American Psychological Association for decades that intelligence can be reliably measured by scientifically-designed and properly g-loaded IQ tests. One well-known scientist: “IQ tests are among the most reliable, predictive measures in psychology — one of the field’s crowning achievements. If IQ isn’t a valid concept, no concept in psychology is valid.” But even if that “thing” that IQ tests measure wasn’t general intelligence (“g”), you would still want lots of it because it predicts (often better than any other known variable) how well someone is going to perform in so many different life outcomes. General intelligence (g) has been found in hundreds of studies to correlate positively with desirable traits such as:
— Financial success — Income and socioeconomic status — Occupational level attained — Educational attainment — Academic achievement
Higher general intelligence (g) has also been positively associated with:
— Job and job-training performance (“Meta-analyses show that [IQ]… is the best single predictor of on-the-job performance.”) — Level of cooperativeness among individuals.
The graph shows the correlations of IQ with different aspects of job performance. You can see that the correlation is stronger the more cognitively complex the occupation.
Other positive associations with higher IQ include:
— Lower rates of most mental illnesses — Better physical health — Lower risk of death from most causes — Longevity On the other hand lower IQ has been consistently found to correlate with undesirable traits such as:
— Anti-social and violent behavior — Criminality — Impulsivity — Divorce — Automobile accidents
Bottom line: Higher IQ tends to correlate positively with desirable traits, and lower IQ with undesirable traits.
So a stock talking point that people not familiar with IQ research use — “IQ tests only measure your ability to take an IQ test” — is empty unscientific rhetoric. IQ clearly measures SOMETHING valuable, and is in some cases the best known predictor of a life outcome.
Cultural fairness
Source: TW
PART 22: Cultural Fairness of IQ Tests Image Some people have argued that the “cultural specificity” of intelligence makes IQ tests biased towards the environments in which they were designed, and therefore a particular test developed for a “white” culture may not accurately measure intelligence in a non-white one. Although this argument isn’t really made that much anymore in published research because it hasn’t held up in studies, it still occasionally appears in some college textbooks, according a recent review of intelligence instruction materials. A survey of intelligence experts conducted in 2008 found that only 13% of them believed that IQ tests are racially or culturally biased.
The evidence against bias has continued to grow since then; that percentage is very likely even less now. Efforts to create tests in which black performance would improve produced instruments which contained very little g-loading (i.e., didn’t really measure general intelligence).
A 2000 review of the literature concluded that “the issue of test bias is scientifically dead.” A recent study found that general intelligence (“g”) is a “universal phenomenon” found in all 31 non-Western nations investigated. Therefore, “it is theoretically possible to conduct [culturally appropriate] cross-cultural research on intelligence.” Another analysis found that the most common cognitive testing construct (i.e., theoretical framework that underlies IQ tests) was found to have no apparent cultural bias. It’s also difficult to imagine IQ tests developed in the West as culturally biased when northeast Asians score higher on them than whites. East Asians average higher scores than whites both in the US and in Asia, and even on tests specifically designed for “whites”. Finally, there is an simple and reliable way to gauge general intelligence that is almost indisputably culture-free — that test being reaction time — and this type of testing shows the same differences between races that standard IQ tests show. This is because “reaction time measures the neurophysiological efficiency of the brain’s capacity to process information accurately — the same ability measured by intelligence tests.” A combination of reaction time measures produces a multiple correlation of .67, about the same magnitude as the correlation between reasoning ability and vocabulary — so it’s a good indicator of IQ. Unsurprisingly then, various types of reaction time testing in multiple studies produced consistent results in the familiar descending order: Northeast Asians fastest (both in US and Asia), whites intermediate, and blacks slowest.
Stereotype threat
PART 23: Rebutting the Environmentalist Claim of Stereotype Threat
One of the most popular environmentalist arguments in the race and IQ debate has been stereotype threat.
This theory posits that blacks aren’t less smart on average, they just perform worse on IQ tests because they’re conforming to existing stereotypes about them. SIDE NOTE: Group stereotypes are actually typically accurate. “Stereotype accuracy” is one of the most powerful and replicable effects in all of social psychology. The main problem with stereotype threat is that “the effects either disappear or reverse altogether in the highly powered large scale studies.” And it flunks meta-analyses. “Overall, results indicate that the size of the stereotype threat effect that can be experienced on tests of cognitive ability in operational scenarios such as college admissions tests and employment testing may range from negligible to small.” “At best, stereotype threat is mildly interesting laboratory effect. In real-world testing situations, its effect is probably subtle — possibly zero.”
Laboratory studies “often do not resemble real testing situations.” Another problem is that some studies contain “stereotype reminders,” i.e., they hint at the existence of racial stereotypes in their design. Unsurprisingly, “studies that have strong stereotype reminders are much more likely to produce evidence for stereotype threat.” If stereotype threat was a valid explanation for black performance on IQ tests, we might expect blacks to lack confidence in their intellectual abilities. In fact, the opposite is true: Blacks score higher than whites on measures of academic self-esteem… … and are more likely to assess themselves as being smarter than average or “among the very brightest.” The replication crisis (i.e., findings not being reproduced in later studies) and the publication bias crisis (e.g., non-replicated studies not being published) in psychology have also played a role in keeping stereotype threat theory alive. So if stereotype threat effects among African-Americans have been mostly obtained in experimental studies with small unrepresentative samples, and don’t hold up in larger more rigorous studies, why is it still being taught? Heterodox Academy answers the question: “Stereotype threat, of course, is a great [egalitarian] tool… it is professionally risky to challenge ideas that serve egalitarian rhetoric.” Stereotype threat comprises much of what’s left in the environmentalist’s toolbox for explaining racial IQ gaps — so if a scientist exposes its problems, that effort comes with some social and professional risk. Finally: Even if stereotype threat was real and had explanatory power, it would only explain a small portion of the black-white IQ gap. Across studies it produces “an average decrease of only 3 IQ points for African-Americans,” leaving 12 points of the gap unaccounted for.
Shared variables
PART 24: Shared and Non-Shared Environmental Variables
So, if environmentalists don’t believe genetic differences between races influence the IQ gaps between them, what do they think causes them?
They attribute the gaps to environmental factors like nutrition, education, poverty, and health care. I’ll discuss those specific factors in my next thread, but today I’ll talk about the two broad types of environments scientists focus on when investigating individual differences in IQ — “shared” and “non-shared.” The “shared” environment relates to the environment shared by siblings reared in the same family (e.g., parent’s socioeconomic status and child-rearing practices), and the “non-shared” environment relates to the environment unique to an individual. It is now widely accepted that the influence of the “shared” environment becomes negligible in adulthood, perhaps disappearing altogether — which means, for example, parental child-rearing practices ultimately have almost nothing to do with IQ variation.
Scientists have found that the IQ correlation between adopted siblings (genetically unrelated people raised together) falls to zero in adulthood — suggesting shared childhood environment has no impact on intelligence in adulthood.
As the effects of the “shared” environment on human intelligence wear off in adulthood, the effect of genes increases (called the “Wilson Effect”) — observe on the graph how the shared environment ultimately drops out of the picture.
A study of Dutch children showed an even earlier drop-off of the role of the shared environment, to zero, by around age 12.
But even before adulthood, while a child is still in the home, “family and parenting characteristics are not significant contributors to variation in IQ scores.” The other type of environment — “non-shared” — is not well-understood, and scientists have been unsuccessful at decisively pinning these variables down.
It is believed, however, that they consist of “many experiences of small effect.” Some scientists refer to the non-shared environment as a “catch-all bin” into which to throw any variable neither associated with genes nor the shared environment — like friends, teachers, role models, diseases and disabilities, and emotionally arousing events in childhood. “[W]e have learned in recent years that [non-shared environmental effects] are mostly random [and] unsystematic… which means that we cannot do much about them.” In other words, they may largely be “chance effects in development.” Along these lines, there is a growing school of thought that perhaps much (if not most) of the non-shared environmental influence occurs in the pre-natal environment of the womb — that is, in a biological, genes-influenced space before birth. Because “intrinsic stochasticity” (i.e., randomness) of molecular processes underlies individual development, “at the critical stages of ontogeny, even minor fluctuations in gene expression or gene-product functioning can remarkably affect [IQ].” To translate: The molecular chaos involved in fetal development (sometimes referred to as “developmental noise”) — over which we have little or no control — may just be the most important NON-GENETIC influence on individual variation in intelligence. So, maybe the reason scientists haven’t been able to identify the specific non-shared environmental factors is because most of those that are actually influencing cognitive development may be occurring haphazardly INSIDE the human body at the molecular level before birth. Another way of looking at this: If genes plus random pre-natal molecular processes are almost entirely responsible for our cognitive development — both of these being “internal” processes — then little room is left for “oppressive social forces” to explain IQ differences. One more point: Some of our environmental influences are genetically influenced because “we select, modify and create environments correlated with our genetic behavioral propensities.” “More than 150 papers… have been published, consistently showing significant genetic influence on environmental measures, extending the findings from family environments to neighborhood, school, and work environments.” Steven Pinker: “[Genes] may play an even greater role than we imagined, because they may account for the ’non-shared’ portion of environmental factors (which is the biggest portion of ’nurture’).”
Bottom line: Genetic and biological influences extend even into the “environment.”