Identifying children in the Colorado Adoption Project at risk for conduct disorder

Clustering techniques were used to identify a subsample of young adopted and nonadopted children in the Colorado Adoption Project at risk for conduct disorder. Although data from both boys and girls were analyzed, a cluster of girls large enough for subsequent statistical analysis could not be identified; therefore, results are reported for boys only. Identifying measures were selected based on the DSM-III-R diagnostic criteria. Cluster analyses confirmed the existence of a small group of boys who appeared to be significantly at risk. Subsequent parental and teacher ratings of these children verified the stability over time of the classification. The poor conduct group was significantly associated with difficult temperament in infancy, with poor conduct on the part of parents when they were youths, and with high achievement orientation in the home environment.

Quantitative genetic analysis of IQ development in young children: multivariate multiple regression with orthogonal polynomials

The study of psychological development has recently benefited from innovative analytic methods for estimating and examining the correlates of individual growth curves. These methods are more consistent with a conceptualization of development as an ongoing, continuous process, rather than as increases or decreases in a trait between two discrete time points. Recent developmental behavior genetic models have focused on continuity and change in the genetic and environmental influences underlying phenotypes. In contrast, we present a model for genetic and environmental influences on phenotypic development per se. In this model, we adapted multiple regression methods developed for twin designs (DeFries and Fulker, 1985) to a parent-offspring adoption design and to a multivariate framework in which repeated measurements are decomposed into orthogonal polynomial trends. We applied these analyses to the development of IQ during infancy and early childhood using parent-offspring data from adoptive and nonadoptive families in the Colorado Adoption Project. The results suggested familial environmental influences on children's mean IQ for ages 1-4 but environmental influences specific to fathers' cognitive ability on children's IQ development. We also discuss advantages and disadvantages of the multivariate multiple regression method for studying genetic and environmental influences on development.

Differential heritability across levels of cognitive ability

Differences in heritability and shared environmentality across levels of cognitive ability were assessed in a sample of 264 twin pairs tested at 1 year of age and in subsets tested at 2 and 3 years. Using an extension of the DF multiple regression methodology for analyzing twin data, no evidence was found for a linear or quadratic effect of level of cognitive ability on estimates of heritability or shared environmentality.

Parent-offspring and sibling adoption analyses of parental ratings of temperament in infancy and childhood

A first step toward understanding the etiology of personality is to investigate the relative impact of genetic and environmental factors using twin and adoption designs. Twin studies of infants and young children indicate substantial genetic influence for parental ratings of temperament in the preschool years. Adoption studies, however, have not previously been reported during the early years of life. We present parent-offspring comparisons for temperament (emotionality, activity, sociability, and impulsivity) for adopted and nonadopted children yearly from 1 to 7 years of age and their biological, adoptive, and nonadoptive parents. Also presented are correlations for adoptive and nonadoptive siblings when each child was 1, 2, 3, and 4 years of age. In contrast with twin results, little evidence is found for genetic influence. The average correlation between biological parents and their adopted-away children for data averaged over the 7 years is only .03. Similarly, the average parent-offspring correlation in nonadoptive families (.08) is no greater than in adoptive families (.12). Results for nonadoptive and adoptive siblings also indicate little genetic influence. The difference between the twin and adoption results may be due to environmental effects or to nonadditive genetic variance.

A LISREL 8 model with constrained parameters for twin and adoptive families

Constrained optimization recently has been implemented in the LISREL software package, allowing formulation of parent-offspring transmission models in a simple and efficient manner. A reverse path model of parental transmission is described within the LISREL framework for application to twin and/or adoptive family data. The model incorporates genetic and environmental parameter constraints arising from assortative mating and cultural transmission. An illustration of the LISREL model is given using measures obtained from twins and parents involving fear of social criticism.

Longitudinal prediction of specific cognitive abilities from infant novelty preference

A test of visual novelty preference, the Fagan Test of Infant Intelligence, was administered to a group of 113 full-term infants from the Colorado Adoption Project at 5 and 7 months of age. The infants were followed longitudinally and the Bayley scales were administered at 12 and 24 months, the Sequenced Inventory of Communication Development at 24 and 36 months, and the Stanford-Binet and the Colorado Specific Cognitive Abilities Test at 36 months. 1 novelty preference score was obtained for each infant by averaging across the 2 test ages. Novelty preference correlated significantly with 36-month Binet IQ, the first unrotated principle component from the cognitive battery, and the 24-month Bayley MDI score. Novelty preference was also compared to specific abilities at all 3 follow-up ages; all of the specific abilities were significantly related to novelty preference, with the exceptions of 12- and 24-month Imitation and 36-month Perceptual Speed. Partial correlations suggest that novelty preference predicts language and memory independent of IQ. Overall, the results indicate that novelty preference during the first year of life not only predicts later IQ but may also reflect specific cognitive processes.

A simple method of model fitting for adoption data

Traditional models used with adoption data often make strong assumptions concerning the nature of genetic transmission and assortative mating. A simple model is presented which avoids these assumptions. The model is linearized and, thus, has the further advantage that it can be used with standard statistical packages such as LISREL or EQS. The model allows tests of the internal consistency of the data, in addition to tests of the relative strength of genetic and environmental transmission parameters. To illustrate the model, measures of general cognitive ability in parents and their 7-year-old children from the Colorado Adoption Project (CAP) were fit to the model using the LISREL program. This relatively simple model may be expanded to incorporate more complex designs involving multiple measures or siblings. Although the model will not always allow constraints on the parameter estimates in more complex models, it offers a quick, flexible method for initial exploration of adoption data.

Twin temperament during the transition from infancy to early childhood

A multiple-regression model for the analysis of twin data in which one twin's score is predicted from that of its co-twin, the coefficient of relationship, and the interaction term provides direct estimates of heritability (h2) and the proportion of variance due to shared environmental influences (c2). Several multiple-regression models were fitted to parental ratings of infant and toddler temperament for 306 pairs of twins. High estimates of h2 were found for eight of the nine temperament scales, and three scales showed significant effects of c2. No differential heritability was found between males and females across the infant and toddler years.

Quantitative genetic analysis of longitudinal trends in adoption designs with application to IQ in the Colorado Adoption Project

A factor model is presented that provides for either multivariate or developmental specification of longitudinal genetic and environmental effects in the presence of assortative mating and cultural transmission. Delta path methods are employed for the treatment of assortative mating and selective placement effects. The proportions of genetic and environmental variance and covariance attributable to assortative mating and cultural transmission are modeled explicitly. The model was applied to cognitive ability data on 493 families in the Colorado Adoption Project by means of maximum-likelihood pedigree analysis. A test of the assumption of multivariate normality of error provided an additional model criterion beyond the log-likelihood ratio statistic. No significant effects were found for cultural transmission, genetic-environmental covariance, or selective placement. The results suggest that the phenotypic stability of IQ during early childhood is largely, if not entirely, genetic in origin and that these longitudinal genetic effects can be represented most parsimoniously in the form of developmental transmission.

Multivariate path analysis of specific cognitive abilities in the Colorado Adoption Project: conditional path model of assortative mating

A multivariate path model of genetic and environmental transmission employing a conditional path representation of assortative mating was fitted to specific cognitive abilities data from the Colorado Adoption Project and evaluated using a maximum-likelihood estimation procedure. In agreement with results obtained from a previous analysis of a smaller data set, significant genetic covariation among the cognitive variables was indicated and evidence for a general genetic factor was found. However, cultural transmission parameters are nonsignificant and environmental correlations among the measures are relatively small.

The effects of social homogamy and cultural inheritance on the covariances of twins and their parents: a LISREL model

In general, models involving parent-offspring transmission are not possible to specify in LISREL due to the complex constraints implied by assortative mating and genotype-environmental correlation. In this short note we describe one simple model of resemblance among twins and their parents which can be accommodated in the LISREL specification due to the strong assumption of social homogamy. The specification is described fully and checked for identification, and its power to resolve different parameters is briefly examined.

Fitting genetic models with LISREL: hypothesis testing

A brief introduction to the mathematical theory involved in model fitting is provided. The properties of maximum-likelihood estimates are described, and their advantages in fitting structural models are given. Identification of models is considered. Standard errors of parameter estimates are compared with the use of likelihood-ratio (L-R) statistics. For structural modeling, L-R tests are invariant to parameter transformation and give robust tests of significance. Some guidelines for fitting models to data collected from twins are given, with discussion of the relative merits of parsimony and data description.

Testing structural equation models for twin data using LISREL

Simple genetic models can be fitted to twin data using software packages such as LISREL (Jöreskog and Sörbom, 1986a). After discussion of data preparation and routine checks on possible violation of assumptions of the twin method, we illustrate univariate, bivariate, and multivariate genetic models which can be tested in cross-sectional twin data using LISREL. These include models for cohort or cohabitation effects, genotype x sex interaction, and certain types of genotype x environment interaction and genotype-environment correlation.

Genetic influence on general mental ability increases between infancy and middle childhood

Adoption studies can provide direct evidence for the independent effects of family environment and heredity that are always confounded in intact nuclear families. When children are separated from their biological mothers shortly after birth and placed nonselectively in adoptive homes, adoptive-parent/adopted-child resemblance can be ascribed to cultural transmission, whereas biological-parent/adopted-child similarities are due to heritable factors. Furthermore, a longitudinal adoption study facilitates examination of changes in these two main sources of variation during development. The Colorado Adoption Project is the first large-scale longitudinal adoption study of behavioural development and was initiated in 1975. Data were collected from biological parents of 245 adopted children, the adoptive parents and parents of 245 matched nonadopted children. The children have subsequently been tested at 1, 2, 3 and 4 years of age, and at the end of their first year in primary school (average age, 7.4 years). The number of subjects tested is now adequate for analysis of data over 7 years. The results provide conclusive evidence for increasing heritable variation of general mental ability, ranging from 9% at 1 year of age to 36% at 7 years.

Multiple regression analysis of twin data: etiology of deviant scores versus individual differences

The multiple regression analysis of twin data in which a cotwin's score is predicted from a proband's score and the coefficient of relationship (the basic model) provides a statistically powerful test of genetic etiology. When an augmented model that also contains an interaction term is fitted to the same data set, direct estimates of heritability (h2) and the proportion of variance due to shared environmental influences (c2) are obtained. A simple transformation of selected twin data prior to regression analysis facilitates direct estimates of h2g (an index of the extent to which the difference between the mean of probands and that of the unselected population is heritable) and a test of the hypothesis that the etiology of deviant scores differs from that of variation within the normal range.

Evidence for a genetic aetiology in reading disability of twins

Reading disability (dyslexia) is a major social, educational, and mental health problem. Although estimates of prevalence vary, up to 10-15% of school-age children have severe reading deficits in spite of average intelligence and adequate educational opportunity. That reading disability may have a constitutional basis has long been recognized, and results of twin and family studies suggest that one or more of its forms may be heritable; however, definitive evidence for a genetic aetiology has not been reported. Establishing a heritable basis for reading disability could suggest possible causes, give improved risk estimates, facilitate early diagnosis, and provide validity tests for ostensible subtypes. In this report, we apply a recently developed multiple regression analysis to data collected from a sample of 64 pairs of identical twins and 55 pairs of fraternal twins, in which at least one member of the pairs is reading disabled, and present evidence for a significant genetic aetiology.

Path analysis of seven fear factors in adult twin and sibling pairs and their parents

A multivariate path model of genetic and environmental transmission, used to derive expected covariances among adult twin and sibling pairs and their parents, was fitted to fear factor data from 250 twin families and 91 sibling families, with the use of a maximum-likelihood estimation procedure. The full model, with 259 free parameters, provides for parental cultural transmission, common twin and sibling environment, genotype-environment correlation, direct assortative mating, and genetic and environmental correlations. Significant effects were indicated for heritable transmission of common fears and phobias, and a single genetic factor accounted for most of the genetic covariance among the traits. Moderately high levels of common twin environment and a small effect for direct isomorphic marital assortment were also found. There was no evidence for cultural transmission, genotype-environment correlation, or common sibling environment.

Longitudinal stability of cognitive ability from infancy to early childhood: genetic and environmental etiologies

A path model of genetic and shared family environmental transmission was fitted to general cognitive ability data from 1-, 2-, 3-, and 4-year-old adopted and nonadopted children and their parents in order to assess the etiology of longitudinal stability from infancy to early childhood. Stability across years is moderate and is due mainly to influences not predicted by parental IQ. Results of the present study, in conjunction with those of previous twin studies, suggest substantial genetic stability from infancy and early childhood to adulthood.

Altruism and aggression: the heritability of individual differences

Five questionnaires measuring altruistic and aggressive tendencies were completed by 573 adult twin pairs of both sexes from the University of London Institute of Psychiatry Volunteer Twin Register. The questionnaires measured altruism, empathy, nurturance, aggressiveness, and assertiveness. The intraclass correlations for the five scales, respectively, were .53, .54, .49, .40, and .52 for 296 monozygotic pairs, and .25, .20, .14, .04, and .20 for 179 same-sex dizygotic pairs, resulting in broad heritability estimates of 56%, 68%, 70%, 72%, and 64%. Additional analyses, using maximum-likelihood model-fitting, revealed approximately 50% of the variance on each scale to be associated with genetic effects, virtually 0% with the twins' common environment, and the remaining 50% with each twins' specific environment and/or error associated with the test. Correcting for the unreliability in the tests raised the maximum-likelihood heritabilities to approximately 60%. Age and sex differences were also found: altruism increased over the age span from 19 to 60, whereas aggressiveness decreased, and, at each age, women had higher scores than men on altruism and lower scores on aggressiveness.

Multiple regression analysis of twin data obtained from selected samples

The multiple regression analysis of twin data in which a cotwin's score is predicted from that of a proband (the member of a twin pair selected because of a deviant score) and the coefficient of relationship provides a powerful test of genetic etiology (DeFries and Fulker: Behav Genet 15:467-473, 1985). Moreover, when an augmented model containing an interaction term is fitted to the same data set, direct estimates of heritability (h2) and the proportion of variance owing to shared environmental influences (c2) are also obtained. In the present paper, the expected partial regression coefficients estimated from these models are derived, and the flexibility of the general approach is illustrated. An extended model is formulated for the analysis of data from combined samples of affected and control twin pairs that yields tests for differential h2 and c2 in the two groups as well as pooled estimates of these parameters. The application of these models is illustrated by an analysis of data from reading-disabled and control twin pairs. Because of the ease, flexibility, and utility of the multiple regression analysis of twin data, it is an appealing alternative to more traditional model-fitting approaches.

Altruism and aggression: The heritability of individual differences

Five questionnaires measuring altruistic and aggressive tendencies were completed by 573 adult twin pairs of both sexes from the University of London Institute of Psychiatry Volunteer Twin Register. The questionnaires measured altruism, empathy, nurturance, aggressiveness, and assertiveness. The intraclass correlations for the five scales, respectively, were .53, .54, .49, .40, and .52 for 296 monozygotic pairs, and .25, .20, .14, .04, and .20 for 179 same-sex dizygotic pairs, resulting in broad heritability estimates of 56%, 68%, 70%, 72%, and 64%. Additional analyses, using maximum-likelihood model-fitting, revealed approximately 50% of the variance on each scale to be associated with genetic effects, virtually 0% with the twins' common environment, and the remaining 50% with each twins' specific environment and/or error associated with the test. Correcting for the unreliability in the tests raised the maximum-likelihood heritabilities to approximately 60%. Age and sex differences were also found: altruism increased over the age span from 19 to 60, whereas aggressiveness decreased, and, at each age, women had higher scores than men on altruism and lower scores on aggressiveness.