Parental Education and Child’s Verbal IQ in Adoptive and Biological Families in the National Longitudinal Study of Adolescent Health

Michelle Neiss1 and David C. Rowe1,2

Received 21 June 1999—Final 12 Oct. 2000

This study compared adoptive children and matched, biological children to estimate the genetic and environmental effect of years of mothers’ and fathers’ education on children’s verbal in- telligence (VIQ), as assessed by knowledge of vocabulary words. Adoptive and biological ado- lescent children in the National Longitudinal Study of Adolescent Health (Add Health) were matched for sex, age, parental education, and ethnicity. The adolescents all resided with two parents. Structural equation modeling was employed using Mx to estimate the genetic and trans- missible environmental components of the correlation between parental education and children’s VIQ. The mother–child and father–child correlations in biological families were .41 and .36, re- spectively, vs .16 and .18 in adoptive families. As suggested by these correlations, both genetic and shared environmental influences were statistically significant in the Mx models. We con- clude that parental education exerts a modest shared environmental effect, explaining no more than 3 to 4% of the variation in verbal intelligence.

KEY WORDS: Adoption study; verbal IQ; heritability; shared environment.


Behavioral geneticists have studied the influence of heredity on IQ for many decades, with most of the ev- idence pointing to a substantial genetic component to general cognitive abilities (Plomin and Petrill, 1997). Numerous studies with diverse samples and various methodologies have converged on this result. For ex- ample, Scarr and Weinberg (1983), in a U.S. sample, found that adopted children’s IQ scores were more closely correlated with the educational level of their biological parents than with the IQ scores of their adop- tive parents (IQs for biological parents of the adoptees were unavailable). Correcting for selective placement, Scarr and Weinberg estimated the heritability of IQ to range from 40 to 70% in this sample. Boomsma et al. (1998), looking at twins in Holland, estimated the her-


210033, University of Arizona, Tucson, Arizona 85721.

2 To whom correspondence should be addressed. e-mail: dcr091@

itability of IQ to be about 62%. Pal et al. (1997) found that genetic influences accounted for 81% of the vari- ance in intelligence in a sample of rural Indian ado- lescents. Looking at adult identical twins reared apart, Bouchard et al. (1990) estimated the broad heritabil- ity of IQ to be about 70%. In the Colorado Adoption Study (Plomin et al., 1997), genetic influences on ver- bal ability became stronger as children grew up from infancy to adolescence. In an analysis of kinship cor- relations on IQ from multiple studies, Chipuer et al. (1990) found that additive and nonadditive genetic variance together accounted for 51% of the variance in IQ scores.

Although these studies arrive at no single-point estimate of heritability, it is certain that IQ demon- strates a high degree of genetic variation. Genes, how- ever, do not necessarily account for all observed vari- ance in intelligence. Family environmental factors, such as parental socioeconomic class, may also play a role. Scarr and Weinberg (1983) found that genet- ically unrelated young siblings were as similar to each

Division of Family Studies and Human Development, Campus Box


0001-8244/00/1100-0487$18.00/0 © 2000 Plenum Publishing Corporation


Neiss and Rowe

other as biological siblings. Similarity between adopted children and their adoptive siblings suggests that shared environmental effects do contribute to individual differences in intelligence in childhood. Resemblance between adopted children and their sib- lings, however, generally decreases with age. For ex- ample, in the Scarr and Weinberg (1983) study, the IQ correlation between unrelated siblings among older adolescents was 0, whereas adolescent biological sib- lings were similar to one another (r