Abusing Heritability: The Jensen Contradiction

Recently, there has been a major surge in race pseudoscience activism.

The election of Donald Trump has given extremists a lot of confidence, both in violent demonstrations and on social media websites like Twitter. They push pseudoscientific bigotry, yet insist that they themselves are oppressed victims of the evil mainstream media.

Yet, they do not offer that much new material in terms of arguments and claims. Instead, they merely recycle the same arguments over and over, despite the fact that they have been completely destroyed by scientific evidence for many years.

Race pseudoscience activists often abuse scientific concepts and models, including heritability and evolution. This is typically done in very obvious ways and it can be refuted by just looking up the definition of the terms involved. Yet this does not deter race pseudoscience activists. Instead, they merely dig their heels and insist that they are being misrepresented.

This article investigates one such case involving heritability and the race pseudoscience defender Arthur Jensen. As we will see, Jensen insists that he was misrepresented, yet in the very next sentence, he commits the exact same error that he is being criticized for.

What is heritability?

Heritability is one of the most misunderstood concepts in biology. As explained in The Widespread Abuse of Heritability, heritability is not an estimate of “how genetic” a particular trait is, but rather:

To make the difference clear, consider the following example. We know there are powerful genetic factors involved in having five fingers on each hand, but the heritability is almost zero. This is because the most dominant factor for variation in the number of fingers is the variation in exposure to things like industrial or car accidents. There are genetic conditions (like polydactyly) that also plays a role, but it is very rare in comparison.

How do race pseudoscience activists abuse heritability?

Race pseudoscience activists make three core errors when it comes to the concept of heritability.

First, they mistakenly think it is a measure of how genetic a certain trait is. A high heritability, they think, means that genes are causally very important for a certain trait, whereas environment is causally much less important. This in not true because heritability is about how much the variation in a trait can be explained by variation in genetics. It is not about the trait itself, but variation in the trait. It is also not about direct genetic causation.

For instance, Li et al. (2004) and Visscher (2006) found that the heritability of height differs between countries such as China (0.65) and Australia (0.8). This is because the variation in height is more due to the variation in environment in China (e. g. not everyone has access to sufficient nutrients), whereas the variation in height in Australia is more due to variation in genetics because most people do have access to sufficient food. Lai (2006) provides a popular science summary of this type of research.

Second, they wrongly claim that heritability is informative of between-group differences. Any heritability is compatible with any degree of genetic causation for between-group differences. In fact, it does not even follow from the mistaken belief that heritability is somehow a measure of “how genetic” a certain trait is.

Third, race pseudoscience activists erroneously insists that a high heritability means that it is somehow ineffective to modify environmental factors to produce large and meaningful changes. More directly, they argue that it is not worth the effort to improve schools, housing and other social factors for ethnic minorities. In other words, race pseudoscience is a very convoluted ideology used to justify the status quo. However, we know from adoption studies and studies that send people from very impoverished environments to really good schools that improving the environment still has a lot of beneficial effects (Nisbett et al. 2012). In other words, high heritability does not entail genetic determinism.

For more detailed refutations of the claims made by race pseudoscience activists see articles such as:

• Abusing Heritability: “Libertarian Realist” Edition
• Abusing Heritability: “Libertarian Realist” Edition (Part II)
• The Pseudoscientific Fabrications of Mass Murderer Dylann Roof
• Ethnic Diversity, Psychosis and Cancer
• Modern High-Throughput Genomics Versus Race Realism

The Jensen contradiction

Yet, many race pseudoscience defenders insist that this is a straw man. They insist that race pseudoscience activists do not at all claim that heritability is informative about the causes of between-group differences. One of the most common citations used to attempt to back up this claim is the Jensen (1972) book “Genetics and Education” (p. 29):

In the first place, I had never claimed that the high heritability of intelligence within either or both racial groups was sufficient to prove that mean intelligence within either or both racial groups was sufficient to prove that mean differences between the groups was attributable, in whole or in part, to genetic factors. It is axiomatic in quantitative genetics that within group heritability cannot prove between group heritability. The relationship is one of probability or likelihood, that is, the higher the heritability of a trait within each of two groups, the greater is the likelihood that a mean difference between the groups has a genetic component and the smaller is the likelihood that the group difference is attributable solely to environmental variation.

Note that Jensen first states that he never claimed that within-group heritability was a decisive argument that the cause of between-group differences was genetic in nature. Indeed, Jensen goes so far as to state that this is “axiomatic in quantitative genetics”! Then he directly contradicts himself by insisting that heritability is indeed informative of between-group differences.

Now, Jensen is, of course, mistaken in the second part. We know that heritability is not informative about the causes of between-group differences. We know this because the two concepts are unrelated. Remember, heritability is what proportion of variation in a trait can be explained by genetic variation (in a specific population in a specific environment). It is not a measure of how genetic a trait is.

To drive the point come, here is Visscher, Hill, Wray (2008):

Heritability is informative about the nature of between-group differences

This misconception comes in two forms, and in both cases height and IQ in human populations are good examples. The first misconception is that when the heritability is high, groups that differ greatly in the mean of the trait in question must do so because of genetic differences. The second misconception is that the observation of a shift in the mean of a character overtime (when we can discount changes in gene frequencies) for a trait with high heritability is a paradox. For IQ, a large increase in the mean has been observed in numerous populations, and this phenomenon is called the Flynn effect, after its discoverer. The problem with this suggested paradox is that heritability should not be used to make predictions about mean changes in the population overtime or about differences between groups, because in each individual calculation the heritability is defined for a particular population and says nothing about environments in other populations. White males born in the United States were the tallest in the world in the mid-19th century and about 9 cm taller than Dutch males. At the end of the 20th century, although the height of males in the United States had increased, many European countries had overtaken them and Dutch males are now approximately 5 cm taller than white US males, a trend that is likely to be environmental rather than genetic in origin.

In other words, Jensen firmly believes that heritability is informative of between-group differences. In fact, he insists that heritability is necessary to derive it. He is mistaken. In a hilarious twist of irony, race pseudoscience activists who fight tooth and nail to defend Jensen against the onslaught of scientific evidence are refuted by the words of Jensen himself. In the very same section that they themselves quote in defense.

Conclusion

Race pseudoscience activists abuse scientific terminology in a wide range of ways. One of their favorite terms to mischaracterize is heritability. In reality, it is about how much of the variation in a trait can be explained by the variation in genetics (in a specific population and in an environment), not how “genetic” a certain trait is. They misleadingly assert that heritability tells you something about the causes of between-group differences and even go so far as to wrongly think that a high heritability means that it is not worth modifying the environment.

They do not like to have these facts pointed out to them, so they insist that they never claimed those things in the first place. One way they do this is to cite from a book by race pseudoscience activist Arthur Jensen from the 1970s. In reality, the quote contradicts itself and Jensen does claim that heritability is informative about the causes of between-group differences.

References

Jensen, A. R. (1972). Genetics and Education. Methuen, London.

Lai, C-Q. (2006). How much of human height is genetic and how much is due to nutrition? Scientific American. Accessed: 2012-06-02.

Li MX, Liu PY, Li YM, Qin YJ, Liu YZ, Deng HW. (2004). A major gene model of adult height is suggested in Chinese. J Hum Genet. 49(3):148-53.

Nisbett RE, Aronson J, Blair C, Dickens W, Flynn J, Halpern DF, Turkheimer E. (2012). Intelligence: new findings and theoretical developments. Am Psychol. 67(2):130-59.

Visscher PM, Medland SE, Ferreira MA, Morley KI, Zhu G, Cornes BK, Montgomery GW, Martin NG. (2006). Assumption-free estimation of heritability from genome-wide identity-by-descent sharing between full siblings. PLoS Genet. Mar;2(3):e41.

Visscher, P. M., Hill, W. G., & Wray, N. R. (2008). Heritability in the genomics era — concepts and misconceptions. Nat Rev Genet, 9(4), 255-266.