Yet another paper seems to show that our educational attainment is written in our genes. It claims that “individual differences in educational achievement are substantially due to genetic differences (heritability) and only modestly due to differences between schools and other environmental differences”. It’s been widely reported as a claim that exam grades are down to nature, not nurture.
This is based on comparisons of the figures for identical and non-identical twins. The reason why people use twin studies is because they believe that our personal characteristics are determined genetically, and so that studies of twins will confirm this. That is bad science. You’re supposed to design research so that it can disprove the proposition under test, and twin studies can’t do it. What a twin study could show, in principle, is that where monozygotic (genetically similar) twins are different, that difference cannot be genetic. That is not, however, what any of them try to do.
Genes are not blueprints for later development; the genetic structure (genotype) has to interact with the environment (phenotype). Your height depends on your genes, but it is not determined at birth; if you are starved you may be stunted. (The increase in height in successive generations has to be largely environmental, because the gene pool changes only very slowly.) If there is a genetic link between common genes and attainment, it does not necessarily mean that the level of attainment is determined by genes – it only predicts similar patterns of attainment within a given environment. So it is not possible to show that any level of GCSE scores is down to genes – it’s all about whether people from the same family, with the same home background, with the same school (and often with the same teacher), with the same experience in early years and of the same age will achieve similar results. Put that way, it would be surprising if the results weren’t very similar – the more so because the sample has been selected to exclude twins where one of them is disabled.
Heritability is supposed to examine the extent to which varability in the phenotype is attributable to the genotype. As there is no direct exploration of genes or genetics in most of these studies, what they actually look at is the way that similar characteristics occur in families, and those are attributed to the underlying genetic structure. There are lots of reasons besides genes why educational attainment might run in families – among them culture, lifestyle, language, common experiences, and so forth. The authors suppose that identical and non-identical twins all have similar home backgrounds, so that the differences between them must be down to the issue of whether they’re identical or not. That however depends on the proposition that identical twins are not treated more like each other than non-identical twins are, and that seems implausible. For example, non-identical twins may have gender differences, and children of different genders are liable to be treated differently.
While the study attributes the differences in performance to DNA, DNA was not usually examined. The results are supposed to be about the differences between monozygotic and dizygotic twins, but it makes no serious attempt to determine whether the twins it is studying are either. It states instead that “Zygosity was assessed through a parent questionnaire of physical similarity, which has been shown to be over 95% accurate when compared to DNA testing.” So what the study actually finds is that if parents think their children are really like each other, those children get more similar educational results than they do if their parents think they are different. The authors assume that the explanation for those similarities between twins must be their DNA – and not, for example, whether parents talk to them and treat them in the same way.
Having said that, there is one finding in this paper that brought me up short, and I think it does reflect on policy. The argument is that as the curriculum has become more standardised, less and less variation between results is attributable to the school, and more and more to ‘heritability’ – which really, in this case, means the home background and early years. That has deep implications for educational equality.
The reason why people use twin studies is because they believe that our personal characteristics are determined genetically, and so that studies of twins will confirm this. That is bad science. You’re supposed to design research so that it can disprove the proposition under test, and twin studies can’t do it. What a twin study could show, in principle, is that where monozygotic (genetically similar) twins are different, that difference cannot be genetic. That is not, however, what any of them try to do.
You misunderstand twin studies. The fact is that the twin design is the best available method for studying both the genetic and the non-genetic determinants of behavioral differences. Twin studies can produce heritability estimates (as well as environmentality estimates) anywhere between zero and 100 percent, although either extreme is very rare in practice. There is no a priori assumption of genetic (or environmental) influence.
The reason why the twin method is superior to traditional sociological methods in the study of environmental effects, too, is that twin studies can estimate the effects of environment independently of genetic effects. Traditional sociological studies are hopelessly confounded by genetic effects, and can only rarely be causally informative.
So it is not possible to show that any level of GCSE scores is down to genes – it’s all about whether people from the same family, with the same home background, with the same school (and often with the same teacher), with the same experience in early years and of the same age will achieve similar results. Put that way, it would be surprising if the results weren’t very similar – the more so because the sample has been selected to exclude twins where one of them is disabled.
The logic of the classic twin design is to compare two types of twins, MZ (identical) and DZ (non-identical). Because twins typically share the same home environment and go to the same schools regardless of zygosity (MZ or DZ), any greater similarity of MZ twins compared to DZ twins can be attributed to the fact that MZ twins share 100% of their genes while DZ twins share only 50%.
The fact that a particular twin pair shares the same environment does not mean that the twin method is only informative about the effects of genes and environment is a particular environment. The big twin study of school achievement you referenced is based on a nationally representative sample of twins. What that means is that the results are generalizable to the entire student population of the UK. The effects of the full gamut of home and school environments are included in the parameter estimates. While the environment may not differ within a particular twin pair, it definitely differs between the pairs in study, enabling the estimation of how much those differences between families contribute to school outcomes.
The authors suppose that identical and non-identical twins all have similar home backgrounds, so that the differences between them must be down to the issue of whether they’re identical or not. That however depends on the proposition that identical twins are not treated more like each other than non-identical twins are, and that seems implausible.
The equal environments assumption (EEA) does NOT mean that MZ and DZ twins are treated identically. Rather, it proposes that environmental effects on the phenotype of interest do not differ based on zygosity. For example, MZ twins are known to be more likely to be dressed in identical clothes and to share friends than DZ twins, but it is also known that such things do not make them any more behaviorally similar.
The EEA can be tested in many different ways. For example, DZ twins who have been misidentified as MZ twins by their families are behaviorally no more similar than correctly identified DZ twins. Another test is adoption studies, where it has been found that by adulthood adoptive siblings are generally no more similar than random unrelated individuals.
For example, non-identical twins may have gender differences, and children of different genders are liable to be treated differently.
Sex differences were explicitly modeled in the study. Standardized genetic variance was generally greater in boys.
While the study attributes the differences in performance to DNA, DNA was not usually examined. The results are supposed to be about the differences between monozygotic and dizygotic twins, but it makes no serious attempt to determine whether the twins it is studying are either. It states instead that “Zygosity was assessed through a parent questionnaire of physical similarity, which has been shown to be over 95% accurate when compared to DNA testing.”
It makes no sense to use genetic tests when parental questionnaires provide virtually the same accuracy. Moreover, to the extent that this leads to misclassification, it will attenuate heritability estimates and inflate environmentality estimates.
So what the study actually finds is that if parents think their children are really like each other, those children get more similar educational results than they do if their parents think they are different. The authors assume that the explanation for those similarities between twins must be their DNA – and not, for example, whether parents talk to them and treat them in the same way.
This is really a rather desperate criticism from you. The questionnaire used asks questions about both physical and behavioral similarity, and it has been shown to discriminate extremely reliably between genetically confirmed MZ and DZ twins.
Correction: The questionnaire only asks about physical similarity (which makes sense).
Twin studies make two key assumptions: that similarities in behaviour are attributable to similarities in gentics (twins and familial clusters are deliberately selected because the genetics are similar), and that the environments are so controlled for in twin studies that they are not relevant. Both propositions are debatable – and they assume the truth of position under dispute.
I’m not sure what you’re talking about here. Please elaborate.
A representative sample of twins – if it is representative, which isn’t clarified – would not be representative of the general population.
Yes it would. The phenotypic means and variances of the twins are very similar to the general population. In early twin studies, conducted in the first half the 20th century, there was evidence that twins tended to be more frail than singletons, leading to lower heritability estimates, but this is not observed in modern cohorts.
◦You included my comment about disability, but you haven’t realised how important it is. MZ twins where one is disabled have provably the same genotype but different phenotypes. So why leave them out?
The GCSE study was about the genetic and non-genetic causes of differences in school success among “normal” adolescents. Those diagnosed as disabled were not included because the etiologies of their academic problems are generally qualitatively different from those of healthy adolescents. Severe mental retardation, for example, is usually innate but NON-heritable and NON-familial. It is caused by de novo mutations, chromosomal anomalies, brain trauma, infections, etc.
Had disabled adolescents been included, it would probably have inflated the “non-shared” variance component of the study which is a sort of noise category used in behavioral genetic studies, denoting those influences that are linked neither to heritable genetic variation nor to the shared family environment. However, the effect of including them would not have been big, because, firstly, the disabled constitute a small share of the total population, and, secondly, lots of the disabled are not able to do GCSEs.
The assumption that “environmental effects on the phenotype of interest do not differ based on zygosity” is preposterous.
The assumption made in studies of twins raised together is that causally relevant shared environments do not differ by zygosity. This assumption has been borne out in the many studies that have tested it. Moreover, it is corroborated by study designs that do not rely on this assumption, such as studies of twins raised apart, adoption studies, and GCTA. The opponents of the twin method assume the truth of their position without reviewing the evidence.
More contentiously, it also says, as I pointed out, that it does not matter that all the MZ pairs are of the same gender and half the DZ pairs are of different genders. This is not considered within the paper. Same sex and different sex pairs were sampled but differences between them are not reported in the study. The gender difference that is reported is between girls and boys.
Table 2 in the paper reports intraclass correlations for each sex and zygosity combination.
◦Parental questionnaires do not provide ‘virtually the same accuracy’ as DNA testing. 95% similarility is far too low, because it allows for more type 1 and type 2 errors than the stats can withstand.
Okay, I grant that the study underestimates heritability and overestimates environmental influences because the parental questionnaire is not 100% reliable. However, this effect is small and its direction is opposite to what you appear to think it is.
There is another methodological shortcoming in the study that biases it against detecting genetic influences. The study assumes that the parents of the students got together randomly, without assortment based on ability, education, or other heritable characteristics. This means that any excess similarity of the DZ siblings caused by assortative mating is treated as a purely environmental effect, even though this is clearly not the case.
The format doesn’t lend itself to long correspondence, so with apologies to my anonymous correspondent, I will keep it short. The core point is that twin studies assume the genetic influence they are supposed to prove. Twin studies have intrinsic limitations. They cannot separate the influence of genes from the influence of a common environment, because twins are commonly treated similarly, and identical twins are treated more similarly than non-identical twins. They cannot distinguish genotype from environmental influences. Studies of identical twins reared apart, which might do some of the work, are difficult to obtain and polluted by small numbers, the strenous efforts of adoption agencies to ensure similarity of background and culture and a disturbing history of fraudulent fabrication of data.
There is an obvious alternative, environmental explanation for similarities between twins: these are people whose parents think they are alike, who have the same home background, diet, amenities, culture, exposure to vocabulary, schooling, teaching and so forth. Similarities are not surprising – and the studies strenuously sift out factors, like disability and gender difference, which might do anything to show that twins are not fated to be alike. (Table 2 is part of the problem.) Dismissing these influences as ‘not causally relevant’ is outrageous – and a good illustration of why so many social scientists no longer take these studies seriously. After nearly 150 years of trying, no-one has shown that our genetic structure has any general effect on these kinds of social issue.