Original link: https://headsalon.org/archives/9061.html
【2022-02-20】
@yiqin_fu Kathryn Paige Harden’s book “The Genetic Lottery” summarizes decades of research on the correlation between genes and various human traits. Whether comparing fraternal twins, siblings, all people, the resulting heritability coefficient is much higher than I thought (how many percent of the difference in shape can be explained by genetic differences). Several other conclusions of the author are: 1) Be careful when interpreting coefficients, because there are also social factors. This coefficient will also change with social changes; 2) We don’t know the causal mechanism; 3) After acknowledging genetic differences, we can design mechanisms to help everyone (such as wearing glasses for myopic people).
But a bigger problem still exists-most innate differences cannot be completely removed by a solution like “wearing glasses”; skills rewarded by different eras and societies are completely random in morality.
I also especially want to ask friends who are doing moral philosophy, is anyone rethinking Rawls/Nozick/Friedman’s theory based on these empirical studies? Any faction needs to face the matter of “the genetic coefficient is so high”?
Other conclusions that impress me are: 4) The heritability coefficient is highest when a person is in his twenties, because genes and effort interact; 5) Many non-cognitive abilities have even higher coefficients than cognitive abilities, such as being affected by Willingness to persevere after setbacks, curiosity about new things (of course, these concepts are measured by games in the laboratory, you can say that the games in the laboratory are different from real life).
@whigzhou: If you are not limited by the PC filter, you should know these things. There is nothing new in this book, but it passed the PC filter because of the leftist policy proposition. In addition, Moral philosophers didn’t need to revise Rawls to find their footing either, they lost their footing in revisions after WW2, Spencer and his Victorian counterparts coexisted well with the facts of evolution
@whigzhou: Most of the traits that are closely related to individual achievements have a 50-70% heritability. This fact has long been known in the circle, but many people outside the circle do not know it. They are surprised by this, and even expressed it Unbelievable, it means that the PC filter is really powerful,
But what I find more interesting is that after hearing the figure of 50-70%, many people are glad that they have at least a chance to change the remaining 30-50%. In fact, this is also a big misunderstanding, and this misunderstanding is even more difficult. eliminated, as the rationale behind it is more difficult to explain/understand,
Simply put, most of the remaining 30-50% comes from three random factors:
1) Measurement error: For example, if the same object is measured by different testers, the measured numbers are different, and there must be errors in it. This error is more serious in psychological measurement than in other (such as physiological) fields.
2) Indicator design deviation, that is, the design of a measurement indicator makes it unable to fully and accurately reflect what it originally intended to measure. This problem exists in any indicator, and psychological indicators are of course more serious.
3) Luck component: The so-called luck is something that you cannot intervene and control. In other words, you can’t do anything,
In this way, you can understand why Robert Plomin found that all the individuals they examined shared environmental conditions in their growth (such as whether they shared families, whether they shared parents, whether they shared schools, whether they shared a certain educational model, whether they shared a certain wealth) Level…), all have no contribution to explaining personality differences, that is to say, including these factors will not change the proportion of unexplained part of personality differences at all,
In layman’s terms, it is futile to try to improve children’s personality/intelligence development by changing these factors, (of course, this is by no means to say that doing these things is meaningless, they may have other effects, such as making children happier , more knowledgeable, more educated, more tasteful… these can be very meaningful, but they cannot change their personality and intelligence),
Of course, don’t be too desperate. It’s not that you can’t do anything about personality/mental development. One of the flaws of those measurements is that their sampling objects are basically ordinary normal families in the United States, so some extreme conditions are automatically excluded. And some extreme conditions do have a significant impact,
For example, lead pollution and malnutrition above a certain level will obviously affect intellectual development, severe abuse in early childhood may also affect certain personality components, malnutrition and calcium deficiency will certainly affect height…
So, there is at least one thing parents can do for their children’s physical/personality/intellectual development: help them avoid those vicious extreme conditions,
@疑谢: If the environment below the average has an impact, does the environment far above the average (such as the top 0.1% of households) also improve?
@whigzhou: 1) Vicious extreme conditions are far more than just *below average*, 2) Positive extreme conditions are useless, just like you can’t raise Yao Ming if you drink tons and tons of milk and calcium supplements
@windflower188026: So what is the explanation of IQ median regression
@whigzhou: Mean regression is a specific manifestation of luck factors. It can be understood in this way: the deviation from the mean is not only the independent effect of some contributing factors, but also the effect of their specific combination, and the lucky combination may be in the process of sexual reproduction. is broken while humans are sexually reproducing creatures
@whigzhou: I’m idle today, so let’s expand on the problem of mean regression again.
Assume that intelligence has five contribution factors, AE, and the contribution rates (r) are: A=3, B=1.2, C=5, D=0.5, E=1, which means that when other conditions are the same, one of them It will increase intelligence by r base points (let’s not worry about what base points mean here),
Assume that these factors are additive, which means that each additional factor is better, but they are not linear additive, which means, for example, that A alone is 3 basis points higher than A alone. B is 1.2 basis points higher, but at the same time (A+B) is not higher (3+1.2=4.2) basis points, but higher or lower than 4.2. If it is higher, it means that A and B have a combined dividend ,
Individuals that deviate greatly from the mean often have a particularly lucky combination (that is, the total contribution of each contributing factor they own contains more combination dividends). The farther the deviation, the higher the luck of the combination. Therefore, in sexual reproduction The greater the combination dividend lost due to the breakage of a specific combination, the greater the range of regression to the mean,
In existing methods, the calculation of heritability is based on the statistical correlation between genetic relationships and individual phenotypes, rather than on the direct understanding of the microscopic mechanism of each contributing factor, which means that, statistically, The combination bonus cannot be separated from the luck component,
For example, A’s father has (A, C, E) among the five factors mentioned above, and his mother has (A, B, D). After reorganization, A has (A, C, D), and (A, C, E ) happened to be a particularly lucky combination, but it was broken during the breeding process, so that, even if intelligence is fully determined by genotype, A’s intelligence value cannot be fully explained by the intelligence value of its parents (because the full explanation requires knowledge of the contributions the microscopic mechanism of action of the factors), while the unexplained part seems to be just luck,
@whigzhou: An additional sentence: If the contribution factors are linearly additive, then the intelligence of a single child cannot be fully explained by the intelligence of the parents, but if the couple has enough children, the average intelligence of the children can be explained by the two of them The intelligence of the children is fully explained, but if the contribution factors are not linearly additive, then no matter how many children are born, the average intelligence of the children cannot be fully explained by the intelligence of the parents
This article is transferred from: https://headsalon.org/archives/9061.html
This site is only for collection, and the copyright belongs to the original author.