“Hereditary and Environmental Sources of Trait Variation and Covariation”, 1972 (; similar):
The twin design for estimating proportions of hereditary and environmental sources of trait variation was presented and applied to a national sample of 806 twin sets who took the National Merit Scholarship Test in 1962. Parental report of differential treatment of their twins was used to test the assumption of equivalent within-family environments by zygosity.
A comparison of the sum of items reflecting differential treatment reported by the parents showed that identical twins are reported to be treated more alike than fraternal twins. Correlations of the treatment difference score with twin differences on the NMSQT and CPI scores showed a small but positive relationship between differential treatment and differences in measured achievement and personality. Within each actual zygosity group, the treatment difference scores of twins whose parents were correct about the zygosity diagnosis were compared to the scores of twins whose parents misdiagnosed them. These results indicated that parental behavior towards their twins is determined largely by the degree of genetic relatedness of their twins. However, the ordering of the treatment difference score means indicated that parental belief about zygosity also determined to some small degree their treatment of their twins. Within each zygosity group, the score differences on the NMSQT and CPI scales of twins correctly and incorrectly diagnosed by their parents were also compared, and the results showed that parental belief about zygosity has a small but consistent relationship to twin differences on measured achievement and personality.
This series of analyses indicated that the assumption of equal between-family environments by zygosity cannot be made, and that the environmental bias is greater for personality measures than for achievement measures. The assumption of equivalent between-family environments by zygosity was also tested, and it was concluded that this assumption does not introduce a serious bias in this sample.
Probable ranges of proportions of trait variance due to heredity, between-family and within-family environment were computed for each measure. Hereditary variation generally accounted for the majority of the variation in the NMSQT scales, and the between-family environmental component was generally larger that the within-family component. The heritability estimates of the CPI scales were quite varied, but in general the within-family environmental component was larger than the between-family component.
A multivariate method by which trait covariation can be partitioned into hereditary and environmental sources was presented and applied to the NMSQT scales. Matrices of cross twin correlations and correlations among twin differences were manipulated to produce hereditary and within and between-family environmental matrices.
The factor structures of these 3 component matrices were compared to the factor structure of the NMSQT. The verbal and math-science factor in the NMSQT were found in the hereditary and the within-family environmental matrices. Only a general factor was apparent in the between-family environmental matrix. This indicated that the 2 factors in the NMSQT are controlled by somewhat different hereditary mechanisms as well as different within-family environmental influences.
[Keywords: genetic correlation, environmental correlation, behavioral genetics, twin study, 1962 National Merit Scholarship sample, NMSQT, CPI, equal environment assumption, mistaken zygosity, personality, academic achievement, factor analysis, thesis]