Airbrushing heritability

D. Wahlsten

DOI: https://doi.org/10.1046/j.1601-1848.2003.00031.x

2003-12-01

Abstract:Genetics and Human Behaviour: The Ethical Context by the Nuffield Council on Bioethics (2002) is the outcome of an extensive review of quantitative genetic studies conducted by a working party of 13 scholars in the United Kingdom that was based on their own reading of the scientific literature as well as interviews and consultations with 40 experts in behavior genetics and opinions submitted by 44 interested organizations and 64 individuals, mostly academics. After reviewing the history of behavior genetics and its relationship with eugenics, the document introduces the reader to the methods of heritability and linkage analysis, touches briefly on animal models and then presents a review focused on the quantitative genetics of intelligence, personality, antisocial behavior and sexual orientation. Much of the document appears to be a one-sided defense of heritability analysis, in which the many critics of that approach have no voice. The document states that: ‘Estimates of heritability and other statistical techniques are useful in understanding the relative contribution of different types of influence.’ (p. xxiii). Finally, there is a lengthy discussion of the practical policy implications of behavior genetic research. The historical review of eugenics in relation to behavior genetics and heritability analysis does not fully confront past or present realities in our field. The Nuffield document acknowledges gross abuses of genetic theory in the 1920s and 1930s but denies that this sordid past means ‘contemporary research on the genetics of behavior is in any sense eugenic or is driven by considerations that could be considered eugenic’ (p. 22). While noting that there has been controversy in recent times concerning the heritability of intelligence in particular, it fails to make a clear connection between the concept of heritability and eugenical selective breeding. On the contrary, it suggests there is no connection in contemporary science. The reader of the Nuffield report might be surprised to learn that the scientific concept of heritability as we know it today was formulated by Lush (1945) expressly for the purpose of selectively breeding farm animals. In a section entitled ‘Practical Applications’ (p. 97), he claimed that it is more efficient to improve the genotype of a breed than its environment, and he argued that the ratio of genetic variance to total phenotypic variance is predictive of the progress of artificial selection (p. 167). The theory of artificial selection was promulgated in the first undergraduate text in behavior genetics by McClearn & DeFries (1973), who asserted that the response to selective breeding R (the difference between offspring mean and mean of the unselected population) is equal to heritability h (narrow sense) multiplied by the selection differential S (the difference between mean of the selected parents and the unselected population mean) or R1⁄4 hS. This formula appeared prominently (p. 281) in the second edition of the text (Plomin et al. 1990), while the third edition (Plomin et al. 1997) moved the topic to the appendix, stating vaguely, without giving the formula, that ‘narrow-sense heritability is particularly interesting in the context of selective breeding studies’ (p. 300). To my knowledge, the R1⁄4 hS equation encapsulates the only practical claim of a heritability coefficient. If one intends to use heritability to anticipate the results of breeding for higher values of a phenotype, it makes a major difference whether heritability is judged to be 30, 40, 50 or 60%. On the other hand, the value of so-called heritability in this broad range of possibilities has no valid implications for the educational or health policy of a nation (Wahlsten 1997a). In order to make valid use of genetic knowledge, we need to know the specific genes that are involved in phenotypic variation, but heritability provides no clue about the number or potencies of relevant loci. About all we can say with confidence is that, if the degree of phenotypic similarity of monozygotic (MZ) and dizygotic (DZ) twins is the same, implying zero heritability, then it is probably not worth doing a linkage study. As Plomin and McGuffin (2003) point out, a number of promising candidate genes pertinent to complex psychiatric disorders have not been confirmed in recent linkage studies and the search for these entities continues. Genes relevant to the normal range of human behavioral variation have been even more difficult to detect, and success is primarily evident for sensory functions (e.g., Kim et al. 2003). The Nuffield report takes note of the arguments of Jensen (1969) and Herrnstein and Murray (1994) concerning the supposedly high heritability of intelligence but does not reveal the links those authors themselves make with contemporary eugenics. While Jensen (1969) argued, wrongly, that high heritability means environmental enrichment programs (e.g., Head Start in the U.S.A.) are doomed to failure, he also warned rhetorically: ‘Is there a danger that current welfare policies, unaided by eugenic foresight, could lead to the genetic enslavement of a substantial segment of Genes, Brain and Behavior (2003) 2: 327–329 Copyright # Blackwell Munksgaard 2003