Friday 27 November 2015

Does social class affect intelligence only in America?


Paper publications impose delays on thought. In compensation, these laggard luminaries of lackadaisical journals claim that the final result is of a much higher standard, polished as the texts are by the sparkling minds of anonymous reviewers. Perhaps so.

On 13 December 2014 I reported from the ISIR conference: In a very big meta-analysis Tim Bates showed that social class interacts with intelligence to some extent in US samples, but not in other parts of the world. It suggests that the much quoted Turkheimer (2003) is something of an outlier in the US funnel plot, but there is a US/rest of world difference, though hard to be sure why, possibly less supportive welfare environment for poor Americans.

Only now in late November 2015 has that paper made its way through the review process, with the the result that I can report to you that in the published paper Elliot Tucker-Drob and Timothy Bates say :

A core hypothesis in developmental theory predicts that genetic influences on intelligence and academic achievement are suppressed under conditions of socioeconomic privation and more fully realized under conditions of socioeconomic advantage: a Gene × Childhood Socioeconomic Status (SES) interaction. Tests of this hypothesis have produced apparently inconsistent results. We performed a meta-analysis of tests of Gene × SES interaction on intelligence and academic-achievement test scores, allowing for stratification by nation (United States vs. non-United States), and we conducted rigorous tests for publication bias and between-studies heterogeneity. In U.S. studies, we found clear support for moderately sized Gene × SES effects. In studies from Western Europe and Australia, where social policies ensure more uniform access to high-quality education and health care, Gene × SES effects were zero or reversed.

Large Cross-National Differences  in Gene × Socioeconomic Status  Interaction on Intelligence. Psychological Science (2015)  1 –12  DOI: 10.1177/0956797615612727

This is an elegant paper, and worth the long wait I have been complaining about. It takes up the Scarr-Salapatek (1971) hypothesis: “IQ scores within advantaged groups will show larger proportions of genetic variance and smaller proportions of environmental variance than IQ scores for disadvantaged groups. Environmental disadvantage is predicated [sic] to reduce the genotype-phenotype correlation in lower-class groups” (p. 1286).

The author’s meta-analysis data set consisted of 43 effect sizes from a total of 24,926 pairs of twins and siblings (approximately 50,000 individuals) participating in 14 independent studies. Data were relatively evenly split across U.S. (18 effect sizes, 8 studies, 10,831 twin or sibling pairs) and non-U.S. (Western Europe and Australia; 25 effect sizes, 6 studies, 14,095 twin or sibling pairs) samples.

Here is the funnel plot of results, with black dots being the US and the red dots the rest:


It is pretty clear that US studies find effects which others do not. So, in the US only, here is the picture:


As you realise, the analysis of variance depends on the circumstances being measured. The authors are probably right to surmise that the US versus The Rest is due to more generous welfare in the rather rich selection of The Rest countries. For example, I doubt we have good data for Brazil or Mexico, but those should show a greater SES effect on g, which would strengthen the interpretation being advanced in this paper.

The authors say:  First, studies from the United States supported a moderately sized Gene × SES interaction on intelligence and academic achievement (a′ = .074; Fig. 1). Second, in studies conducted outside the United States (in Western Europe and Australia), the best estimate for Gene × SES magnitude was very slightly negative and not significantly different from zero. Third, the difference in the estimated magnitude of the Gene × SES effect between the U.S. and the non-U.S. studies was itself significant.

There were no other moderating variables, and no publication bias.

We also replicated the well-established phenomenon that genetic influences on intelligence increase and shared environmental influences on intelligence decrease with childhood age. [] Genes account for considerably more variation in intelligence at both higher ages and in higher U.S. socioeconomic contexts. Indeed, both phenomena may reflect a process of increased and accumulated effects of gene-environment transactions with the increased opportunity that comes with both social class and age.

The results indicate that Gene × SES effects are not uniform but can rather take positive, zero, and even negative values depending on factors that differ at the national level. The finding that low SES was associated with attenuated genetic influence on intelligence in the United States resolves an important debate. The finding that this interaction is observed only in the United States, together with the novel discovery here that the effect may even reverse in sign (The Netherlands), suggests that further research on between-nations variability in the effects of family SES on cognitive development is particularly important. Candidate mechanisms that might underlie such variability include national differences in how concepts of letter and number that underpin literacy and numeracy are imparted, educational quality more broadly, medical and educational access, and macrosocietal characteristics, such as upward social mobility and income support.

This is a very neat result, from a very detailed and carefully argued paper.

As an amusing corollary, as the political Left celebrates the spread of welfare states and seeks to improve social provision in the US, the political Right can take comfort from the proof that, once people have a level playing field of generous benefits, then breeding counts more than ever before.

(I know that the historical picture since 1870 probably favours social spending; that removing benefits would probably have deleterious effects and some good ones; that we may be already detecting dysgenic effects caused by welfare; and that we haven’t identified the candidate mechanisms, but there is amusement in noting how social policies have unintended consequences).


  1. Did they attempt to remove any effects there might be of race, especially for the US? I mean, one obvious difference between the US and the rest of the advanced world might be (in the era that the data came from) the racial composition.

    1. They take into account both the genetic differences and the socioeconomic differences; no matter how you look at the race issue it eventually boils down to some mix of those two factors. And it's not like France and UK are single-race countries, their post-colonial immigration policies have resulted in rather sizeable ethnic minorities of various skin colors.

  2. The sources of the problem are obvious:

    1. Use of children in samples. IQ is still "affected" by shared environment in children.

    2. Racial differences in the U.S. There are much more people in the far low end in the States.

    3. Less variance on the low end (less opportunity for genes to show).

    4. "Double counting": SES is not exogenous, so it's technically not kosher to do these types of studies.