April 14, 2017

The problem with twin studies

Over roughly the last 100 years there has been a debate within the social sciences about the question how large the influence of genes and environments are on human abilities, traits, behaviors, and psychiatric and medical problems. This debate is often referred to as the nature versus nurture debate. It is not only held within the scientific community but also in popular media and by lay people. A central role in these debates is played by the so-called twin studies. Often these studies are viewed as compelling natural experiments which demonstrate the separate influence of genes and environmental factors. And they are often cited as evidence of the large influence of genes on personal characteristics such as intelligence and personality traits. But there is a growing criticism of the validity of these conclusions because there are many problems in those twin studies. A recent book by Joseph (2015) explains in detail what the many problems in twin studies are.

Two types of twins, two types of twin studies

There are two types of twins: monozygotic (identical) twins which share 100% of their genes and dizygotic (fraternal) twins which have 50% of their genes in common. There are also two types of twin studies: studies in which the twins were raised in the same family (Reared Together) and studies in which they have been separated and raised in different families (Reared Apart). Especially the latter type of research is often thought to provide strong evidence for the large influence of genes of intelligence and other personal characteristics.

Problems in the Reared Apart studies

In spite of the big influence of the Reared Apart (RA) studies only very few of these studies have been done and, according to Joseph, all of them are problematic in various ways. There were three classical RA studies which were done between 1937 and 1965 and there is a more recent, very influential, study, The Minnesota Study of Twins Reared Apart (MISTRA; Bouchard et al., 1990). Joseph described the many mistakes and problems in these studies. An important problem in the classical studies is that it has now become clear that most of the twins in the studies were in fact only partly separated and thus were partly reared together, had frequent contact with each other, and had intense emotional bonds with each other. The degree of separation was thus much smaller than was suggested. Another general problem in RA studies is that, even when twins are raised in different families, they usually grow up in very similar cultural and socio-economic circumstances and usually also in the same geographic area.

These types of problems also applied to the MISTRA study which, in addition to this, had many statistical and methodological problems. An extra problem was that the MISTRA researchers declared that skeptics about their study carried the burden of proof to prove that the study's conclusions were false yet they did not grant these skeptics access to the raw data of their study in order to meet this demand. Furthermore the MISTRA researchers did not report all found correlations regarding IQ. They did report all correlation results regarding non-IQ factors but the primary focus area of their study was IQ.

Problems in the Reared Together twin studies

The Reared Together research also has problems. A central premise in this type of research is the assumption that twins which have been reared together have been treated the same and have been exposed to the same environmental factors. On the basis of this premise it is concluded that if identical twins are more similar than non-identical twins this must have been caused by genes, because environmental factors were supposed to be the same. The problem is that the equal environment premise is wrong. It has been demonstrated that identical twins are much more exposed to an equal treated than non-identical twins. Within the logic of these researchers, one could thus equally well conclude that the similarity between identical twins is largely determined by environmental factors.

There is yet another problem which needs to be mentioned. All twin studies are correlational. As is well known, correlational studies can never be used to draw conclusions about causality. If genes were identified for intelligence or personality that would be evidence of the causal influence of genes. But decades of research did not succeed in finding those genes.


There is much more to say about this topic (see, for example, here and here). I want to emphasize that I what I did not want to say here is that genes are not important. On the contrary. Genes are extremely important. They play an important role in the evolution of species and they are also essential for the development of each individual. But the idea that the separate influence of genes on characteristics can be unequivocally determined is dubious, let alone that we now would be able to quantify that unique contribution.

The nature versus nurture debate is outdated. We must now speak of the nature via nurture process of development. How we develop depends on the interaction of our genes with our environments. This means that both genes and environmental factors play indispensable roles in all our characteristics. My suggestion is thus to not be to easily convinced when someone suggests to you that intelligence is largely determined by our genes "because this was proven by twins studies."

No comments:

Post a Comment

Enter your email address:

Delivered by FeedBurner