What can we learn from twin studies? A comprehensive evaluation of the equal environments assumption.
Open Access
- Author:
- Felson, Jacob
- Graduate Program:
- Sociology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- July 04, 2009
- Committee Members:
- John David Mccarthy, Dissertation Advisor/Co-Advisor
John David Mccarthy, Committee Chair/Co-Chair
Roger Kent Finke, Committee Member
Alan Booth, Committee Member
Katerina Bodovski, Committee Member - Keywords:
- equal environments assumption
behavior genetics
twin studies - Abstract:
- Most evidence about the effects of genes on behavior comes from twin studies. The classic twin study method compares the similarity of monozygotic (MZ) twins on some trait with the similarity of dizygotic (DZ) twins on that same trait. If the rearing environments of MZ and DZ twins are analogous, then the magnitude of genetic influence on variation in a trait can be estimated based on the extent to which the correlation for MZ twins exceeds the correlation for DZ twins. If the rearing environments of MZ and DZ twins are not analogous, then the estimates of heritability generated from the classic twin study method may be confounded with differences between MZ and DZ twins in the similarity of their rearing environment. It has long been established that MZ twins grow up in a more similar environment than do DZ twins. However, behavior geneticists have generally asserted that the greater degree of similarity in the rearing environments of MZ relative to DZ twins does not bias the results of twin studies. Behavior geneticists claim that the similarity in outcomes is largely unaffected by similarity in rearing environments for MZ and DZ twins. This claim is known as the equal environments assumption (EEA). In this paper, I argue that the empirical support for the validity of the EEA is not as strong as some behavior geneticists have claimed. I reanalyze some of the data on which ritually cited research on EEA is based, and I find that the evidence is more mixed than latter-day interpretations have suggested. I expose some of the flaws in the research that is frequently cited by researchers in support of claims about genetic influence on behavior. Finally, I conduct the most comprehensive evaluation of the equal environments assumption to date. My analysis incorporates a larger and more diverse set of outcome variables than any previous research on the EEA. In my analysis, I also use a more extensive set of controls for environmental similarity than any previous study. Finally, unlike most if not all previous research on EEA, I generate estimates of genetic influence with and without controls for environmental similarity, in order to show the extent of EEA violation in a straightforward manner. The findings are not easily categorized as supporting or undermining the EEA. Some violations of EEA were found for some outcomes, but the bias caused by these violations is likely to be modest in most cases. Researchers should not conduct twin studies without including controls for environmental similarity. However, the findings also indicate that many of the results of twin studies are robust with these controls. The flaws of twin studies are not fatal, but rather seem no worse (and maybe better) than the flaws of the typical causal study that relies on observational data.