RACE, GENES AND INTELLIGENCE
The issue of the genetic influence on intellectual functioning has historically gone hand in hand with the inflammatory issue of racial differences in intelligence. There is some conflict surrounding the term ‘race’ and whether, in fact, it is a scientifically valid entity. The term ‘ethnicity’ has been suggested as a more accurate alternative, but we use the term ‘race’ here to reflect more accurately the categorization used in the published research.
The race–IQ debate
In 1969, a famous article by Arthur Jensen provided a spark that re-lit the race–IQ debate. He commented that a much-lauded programme of early academic intervention for socially disadvantaged children, known as the Headstart Program, had not resulted in any increase in IQ, and that this was likely to be due to the genetic contribution to intelligence. The idea that an important human trait like intelligence might be, in part, genetically determined and – worse – associated with racial characteristics spawned a stream of outrage, with claims of inherent white Caucasian racial superiority (see Gould, 1996, for a critical review). Such claims had been used in association with the availability of intelligence testing to support a discriminatory immigration policy in the United States in the early part of the twentieth century, favouring Anglo-Saxon immigrants over those of other nationalities on the (plainly ludicrous) grounds that the average IQs of the latter were in the feeble-minded range. As recently as the 1970s, William Shockley, a Nobel laureate for physics (and therefore no expert in psychology), advocated a financial incentive scheme where individuals would be paid not to breed, the amount increasing as IQ decreased. Even more recently, Rushton (1997) has claimed evidence for genetically determined differences between races in many behavioural traits – intelligence being the most important. And finally, publication of The Bell Curve (Herrnstein & Murray, 1994) reignited the fuse primed by Jensen’s famous review. Not only did it claim overwhelming support for the idea that race differences in IQ are in part genetically determined, but it implied that nearly all social disadvantage of racial groups can be traced not to societal discrimination, but to inherited differences in IQ. This issue is large and complex enough to warrant the many books and articles devoted to its discussion (for example, see Gould, 1996; Jensen, 1987; Rose, Lewontin & Kamin, 1984) and also to have evoked a strong and unprecedented publ ic statement signed by 52 intelligence researchers in the Wall Street Journal in December 1994, outlining what is known and what cannot reasonably be extrapolated from research on race and IQ. The core of the race arguments rests on two major propositions, both of which we know are not true:
1. The reliably observed difference in mean IQ – amounting to one standard deviation – between black and white Americans is due to an inherent bias of IQ-type tests against minority groups (see figure 13.10). While some tests undoubtedly show some cultural bias, this is neither systematic nor large enough to account for these reliable IQ differences (see Herrnstein & Murray, 1994; Neisser et al., 1996). Most authorities now accept that the differences measured by IQ tests represent real differences in intellectual attainment. What is denied by most authorities in the field, however, is that these differences are genetic in origin. This brings us to the second untrue proposition.
2. Because we now know that individual diff rences in measured IQ have a large genetic component, it is probable that differences in IQ that may be observed between races are genetic in origin too (Herrnstein & Murray, 1994). What we do know is that the heritability of within-population differences is logically independent of between-group differences in means (Block, 1995).
So, for example, when wheat is planted in a field, some genetic strains will produce more than others under the same soil conditions. In other words, there will be individual differences in yield that are attributable to genetic differences between strains of wheat. On the other hand, the average level of production is likely to be very different in different fields – and this is attributable to the large influence of the different soils (environment) in the fields. So the difference between group means (average wheat yields in each field) and the cause of the within-group individual differences (differences between the genetic strains in each field) are logically distinct. To bring this back to people, what this means is that we cannot necessarily infer that reliable differences between human groups (‘yield of wheat in different fields’) are due to similar mechanisms that determine differences within groups (‘yield of wheat in the same field’). It is perfectly plausible, for instance, that differences in IQ between races are due to differences in the typical environment (‘fields’) across these races.
The issue of the genetic influence on intellectual functioning has historically gone hand in hand with the inflammatory issue of racial differences in intelligence. There is some conflict surrounding the term ‘race’ and whether, in fact, it is a scientifically valid entity. The term ‘ethnicity’ has been suggested as a more accurate alternative, but we use the term ‘race’ here to reflect more accurately the categorization used in the published research.
The race–IQ debate
In 1969, a famous article by Arthur Jensen provided a spark that re-lit the race–IQ debate. He commented that a much-lauded programme of early academic intervention for socially disadvantaged children, known as the Headstart Program, had not resulted in any increase in IQ, and that this was likely to be due to the genetic contribution to intelligence. The idea that an important human trait like intelligence might be, in part, genetically determined and – worse – associated with racial characteristics spawned a stream of outrage, with claims of inherent white Caucasian racial superiority (see Gould, 1996, for a critical review). Such claims had been used in association with the availability of intelligence testing to support a discriminatory immigration policy in the United States in the early part of the twentieth century, favouring Anglo-Saxon immigrants over those of other nationalities on the (plainly ludicrous) grounds that the average IQs of the latter were in the feeble-minded range. As recently as the 1970s, William Shockley, a Nobel laureate for physics (and therefore no expert in psychology), advocated a financial incentive scheme where individuals would be paid not to breed, the amount increasing as IQ decreased. Even more recently, Rushton (1997) has claimed evidence for genetically determined differences between races in many behavioural traits – intelligence being the most important. And finally, publication of The Bell Curve (Herrnstein & Murray, 1994) reignited the fuse primed by Jensen’s famous review. Not only did it claim overwhelming support for the idea that race differences in IQ are in part genetically determined, but it implied that nearly all social disadvantage of racial groups can be traced not to societal discrimination, but to inherited differences in IQ. This issue is large and complex enough to warrant the many books and articles devoted to its discussion (for example, see Gould, 1996; Jensen, 1987; Rose, Lewontin & Kamin, 1984) and also to have evoked a strong and unprecedented publ ic statement signed by 52 intelligence researchers in the Wall Street Journal in December 1994, outlining what is known and what cannot reasonably be extrapolated from research on race and IQ. The core of the race arguments rests on two major propositions, both of which we know are not true:
1. The reliably observed difference in mean IQ – amounting to one standard deviation – between black and white Americans is due to an inherent bias of IQ-type tests against minority groups (see figure 13.10). While some tests undoubtedly show some cultural bias, this is neither systematic nor large enough to account for these reliable IQ differences (see Herrnstein & Murray, 1994; Neisser et al., 1996). Most authorities now accept that the differences measured by IQ tests represent real differences in intellectual attainment. What is denied by most authorities in the field, however, is that these differences are genetic in origin. This brings us to the second untrue proposition.
2. Because we now know that individual diff rences in measured IQ have a large genetic component, it is probable that differences in IQ that may be observed between races are genetic in origin too (Herrnstein & Murray, 1994). What we do know is that the heritability of within-population differences is logically independent of between-group differences in means (Block, 1995).
So, for example, when wheat is planted in a field, some genetic strains will produce more than others under the same soil conditions. In other words, there will be individual differences in yield that are attributable to genetic differences between strains of wheat. On the other hand, the average level of production is likely to be very different in different fields – and this is attributable to the large influence of the different soils (environment) in the fields. So the difference between group means (average wheat yields in each field) and the cause of the within-group individual differences (differences between the genetic strains in each field) are logically distinct. To bring this back to people, what this means is that we cannot necessarily infer that reliable differences between human groups (‘yield of wheat in different fields’) are due to similar mechanisms that determine differences within groups (‘yield of wheat in the same field’). It is perfectly plausible, for instance, that differences in IQ between races are due to differences in the typical environment (‘fields’) across these races.
1 comment:
"What we do know is that the heritability of within-population differences is logically independent of between-group differences in means (Block, 1995)."
This statement is incorrect. Within population heritability is mathematically, and therefore logically, related to between population heritability by the function: BGH = WGH(rg(1-rp)/rp(1-rg). Given the function, high within population heritability (h^2) imposes constraints on between population explanation.
For example, if within group h^2 was 100% and the genetic variance was additive, between group differences could not be explained by factors that vary within populations (SES, bad schools, culture of learning, etc). They would have to be explained by X-factors which are unique to one or the other population and which, in the case of the B-W difference have been shown not to exist.
Now, the adult within population h^2 is ~.75. Assuming additive genetics, to find out how much variable environmental differences you would need to explain the approximately 1 SD adult B-W gap you take the difference divided by the square root of the environmentality. In this case the environmental difference needed to explain the gap is 2 SD of IQ affecting environment (1/sqrt.25). Accordingly for an complete environmental explanation African Americans would have to live in a cognitive environment equivalent to the European American lower 2.5 percentile. As they don't, we can conclude one of two things: either the B-W gap is not wholly environmental or the within population heritable component is substantially non additive.
Post a Comment