“Twins: A Natural Experiment… How to Assess Twin Similarity… Findings from Twin Studies… Are Twin Studies Valid?… Adoption Studies: Not by Twins Alone… Minnesota Study of Twins Reared Apart (MISTRA)… Supplemental – Interview with Tom Bouchard, Director of MISTRA…”
Module F: Minnesota Study of Twins Reared Apart (MISTRA)
Module G: Supplemental - Interview with Tom Bouchard, Director of MISTRA
Module A: Twins – A Natural Experiment
Twin studies are based upon the existence of two types of twins – monozygotic (MZ) and dizygotic (DZ).
Monozygotic twins occur because of a single sperm fertilizing a single egg, and within the first two weeks post conception, the developing embryo divides into two identical zygotes i.e. they have 100% genetic overlap, and they grow up together.
Dizygotic twins occur because two sperms independently fertilize two eggs but in one ovulatory cycle i.e. on average 50% genetic overlap, and they also grow up together.
Twin studies take advantage of this natural occurrence to understand whether or not genetics influences individual differences in a trait i.e. they are natural experiments.
categorical i.e. either or traits like whether one has schizophrenia, or whether one has a college degree, whether a marriage ended in divorce etc.
quantitative i.e. distributed along a continuum and are numeric, such as IQ, extraversion and amount of alcohol consumed in a week etc.
For categorical phenotypes, the measure of twin similarity is usually (twin) concordance i.e. the probability a twin has that phenotype given the co-twin has the phenotype.
For quantitative phenotypes, twin similarity is usually measured using a correlation coefficient i.e. an index of the strength of the linear relationship between two quantitative scores.
Second major type of phenotype is quantitative trait like IQ or extraversion, these are traits that are distributed along a continuum so they are numeric
Step-by-step examples of calculation concordance and correlation coefficient:
Concordance: drug abuse – concordance for monozygotic twins (MZ) is 62%, dizygotic twins (DZ) is 53%, and the population in general is 20%.
Correlation coefficient: height – MZ twins has a correlation of 0.92, a very strong correlation whereas dizygotic twin has a correlation of 0.56
One additional interpretation that will be useful for quantitative genetics and molecular genetics is to multiply it by itself i.e. square it to get the r-squared, which is the percentage of the variance accounted for in one variable by another.
For the height example above, take 0.92 multiplied by 0.92, giving us 0.81 or 81%, which means we can account for 81% of the individual differences in height of one twin once we know the other twin’s height.
Three findings on behavioral phenotypes (i.e. traits) that are fairly consistent across studies:
Monozygotic twins (MZ) are more similar than dizygotic twins.
MZ twins are not perfectly similar.
Similarities in twins are not that different from the similarities in physical and medical traits.
Concordance for MZ higher than for DZ twins on various physical and mental disorders/behaviors such as alcoholism, schizophrenia, Alzheimer’s disease, autism, reading disability, Parkinson’s disease, hypertension, breast cancer, heart disease, peptic ulcers etc.
Correlation coefficient for MZ higher than for DZ twins on psychological characteristics such as memory, neuroticism, extraversion, spatial reasoning, scholastic achievement at adolescence, verbal reasoning, processing speed, general intelligence etc.
Greater MZ and DZ similarity too on “improbable” trains like divorce (concordance), religiousness (correlation), and political attitude (correlation); but we should not conclude there are genes for these traits.
Two explanations for all these results:
Genetic factors might have indirect and multiple effects (i.e. pleiotropic) that influence different traits (because the underlying personality characteristics e.g. the neurotransmission systems in brains might be heritable);
Twin studies are wrong – this will be covered in the next module.
Validity of twin studies is based on two key assumptions:
Representativeness: twins are psychosocially like the non-twin population i.e. the findings can be generalized
Equal Environmental Similarity (EES): environmental similarity of MZ twins is not greater than that of DZ twins.
On the first assumption, studies of psychiatric illnesses, personality and cognitive ability for twins showed they were were similar to those in the general population.
On the second assumption:
Pre-natal environment: the studies that have been done (even though there are not many as it is difficult given the pre-natal environment) do not seem to invalidate the assumption.
Post-natal environment: studies also supported the assumption.
What some studies e.g. on life stress do show is that:
For environmental factors that are externally imposed (such as family-related factors), MZ similarity is not higher than DZ.
For environmental factors that reflect our own behavior (e.g. how we do in school and work), there is greater MZ similarity than DZ similarity.
These suggest that genetic influences on behaviour may be due to the environments we create.
That said, behavioral genetics does not only rely on twin studies (otherwise there might be reservations on the inferences drawn); there are also other studies e.g. adoption studies.
Behavioral genetics does not rely only on twin studies; there are other research design such as adoption studies which also have similar conclusions.
Types of adoption studies:
Adopted siblings: not genetically related but raised in the same home;
Parent offspring: genetic and environmental contributions from one generation to another where the parent and the child have been separated (the latter was adopted by someone).
These adoption studies support the conclusions of the twin studies; they also illustrate that environments can be important.
Module F: Minnesota Study of Twins Reared Apart (MISTRA)
Combine the previous two adoption studies in the form a Reared Apart Twins study – this is a very rare type of study.
One such study is MISTRA, and arguably the most visible behavioral genetic study.
Premise of the study: twins separated at or about the time of birth, and then reunited in adulthood.
Finding: physical and psychological correlations for MZ and DZ were similar for reared apart twins as for reared together twins.
The one exception: IQ – there is still a lot of similarity but not as similar as reared together twins, suggesting the home environments are shaping IQs.
Three general findings from MISTRA (and why it is so influential):
MZ and DZ twins are similar even when reared apart – strengthens the case for genetic influences
Reared apart twins not much less similar than reared together – implies common rearing environment not so influential
Reared apart MZ twins are not psychologically identical – implies environment is important
The first corresponds to A – Additive Genetics, the second to C – Shared Environment, and the third to E – Non-Shared Environment.
Module G: Supplemental – Interview with Tom Bouchard, Director of MISTRA
Some key points from the interview:
Genetic explanations within psychology back in 1979 was mostly negative, particularly true for people who might work on the east coast
On mainstream psychology, the major impact is that students are not taught about behavior genetics
The atmosphere at Minnesota was entirely different as there was an atmosphere that discussed individual differences.
Biggest challenge for MISTRA was identifying the reared apart twins for the study
One phenomena of interest: one of the findings on psychopathology and intelligence is that you can be mentally ill and still be brilliant
Another big issue: measurement in the field of personality is a mess, and no real progress has been made in about 75 years except for finding genetic influence.
Some interesting linkages seem to be coming out of Genome-Wide Association Studies (GWAS) – to be covered later in the course.
