March 23, 2018

expert reaction to study investigating genetic differences in selective and non-selective school students

A study, published in the npj Science of Learning, examines genetic differences between selective and non-selective school students.

This roundup accompanied a briefing.

Dr Ewan Birney, Director of the European Bioinformatics Institute (EMBL-EBI), said:

“I think this is an interesting paper, but it doesn’t get an A+ on the statistics – the sample size is still somewhat too small, and ideally at least two separate educational systems with selection would be studied (say, UK and Denmark). There are stronger statistical techniques which can separate out different hypotheses (e.g., Mendelian Randomisation) which needs larger sample sizes.

“I also think it is less surprising than the headlines might be interpreted; we’ve known that genetics have a reasonable effect on different measures of educational attainment and with a selective system (grammar and private in the UK) unsurprisingly there is implicit selection for genetics. I note the authors themselves stress the weak effects. So the basic facts of this paper should be expected given what we know.

“It is important not to over interpret this paper. These weak effects are unsurprising given educational attainment is a human trait; nearly every human trait has at least some genetics and the biggest effect on academic attainment in this paper is neither genetics nor school – it is something else, very likely each individual.

“Using genetics in broader human traits, such as education, is a useful and sensible thing but we need to be very careful in downstream aspects of this research; the need for replication and constructive criticism in forming robust results, the presentation of the research in the broader press and how it informs policy.”

Prof Gil McVean, Professor of Statistical Genetics, University of Oxford, said:

“In the last few years, population-scale data sets that link genomic data to information on every aspect of human behaviour, biology and disease have shown us that almost every trait is affected by the genes we inherit, and even the genes of our parents that we don’t inherit.  This study builds on previous work that looked at genetic factors that influence educational attainment.  Here, the genetic factors are explored in the context of the type of school attended, though it should be noted that the sample size is relatively small, full data (i.e. also prior attainment, socioeconomic status, etc.) was only available on a subset of the samples (about half) and, perhaps importantly, there are geographic and school-related factors that are not well matched between the schools.

“That there are weak, but statistically significant differences in average genetic propensity for educational attainment between children in different schools is not surprising.  The evidence presented here is not particularly strong, but in the light of previous work, to be expected.  The press release reports this well.

“The authors go on to state in the press release that:  ‘Our study suggests that for educational achievement there appears to be little added benefit from attending selective schools,’ This is a strong claim to make from a very limited data set.

“What they show is that there is a correlation between school type and other factors (including prior attainment, etc.) and GCSE outcome.  It doesn’t mean that attending a school of a particular type isn’t important – it just means you can’t tease out the contribution of the non-genetic factors.  A much bigger and better designed study is needed to provide an answer to this question.  What is clear, is that the contribution of genetics effects is very small in comparison to these other factors.

“It is also worth noting that we have almost no understanding of the mechanism by which genetics affects educational attainment.  We do know that genetic contributions to educational years overlaps with a vast number of other traits, ranging from height to obesity, bipolar disorder to rheumatoid arthritis, depression to snoring.

“In short, I don’t think this paper advances our understanding of the role of selective schools or the biological mechanisms that can impact on educational attainment.”

* ‘Differences in exam performance between pupils attending selective and non-selective schools mirror the genetic differences between them’ by Smith-Woolley et al. published in npj Science of Learning on Friday 23 March.

Declared interests

Dr Ewan Birney: “I consult for Oxford Nanopore and GSK; I am also a non-executive Director of Genomics England.”

Prof Gil McVean: “I have no conflicts of interest, though for complete transparency, I note that I am founder and shareholder in Genomics plc, which is using genomic data to improve healthcare.”