Temperament in middle childhood: A behavioral genetic analysis of fathers' and mothers' reports

Abstract

The genetic and environmental sources of variance in mothers' and fathers' ratings of child temperament in middle childhood were estimated and compared. Parents of 88 MZ twin pairs and 109 same-sex DZ twin pairs completed a temperament questionnaire. For Effortful Control, significant genetic and environmental effects were indicated across mothers' and fathers' ratings, but parent differences were found for the Negative Affectivity factor. When present, sibling contrast effects were not consistent for mothers and fathers. Parental ratings of the Effortful Control factor were best explained by the Biometric model whereas the Negative Affectivity factor was best explained by the Rater Bias model. Overall, mothers' and fathers' ratings yielded similar evidence of genetic and environmental etiology of temperament in middle childhood.

Fig. 1

Sibling interaction model. Twins 1 and 2 are measured temperamental phenotypes for the two twins (e.g., shyness). A, C, E, are latent variables that represent additive genetic influence, shared environmental influence, and non-shared environmental influences, respectively. The curved two-headed arrows represent correlations between the variables they connect. The single-headed arrows a, c, e represents the additive genetic parameter, the shared environmental parameter, and the non-shared environmental parameter, respectively. The b interaction paths refer to the phenotypic interaction or contrast between each twin’s phenotype.

Fig. 2

A bivariate Cholesky decomposition was used to partition the variances of, and covariances between, mothers’ reports and fathers’ reports into genetic and non-genetic components. Separate models were estimated for Effortful Control, Negative Affectivity, and Surgency/Extraversion factors. Latent variables represent overlapping additive genetic effects (A), additive shared environment effects (C), and additive non-shared environment effects including error (E), as well as residual genetic (a), shared environmental (c), and non-shared environmental (e) variance. In this model, the pathways between latent variables representing genetic variance and covariance across twins are set at 1 for monozygotic (MZ, identical) twins and .5 for dizygotic (DZ, fraternal) twins. The pathways for shared environmental variance and covariance across twins are set at 1 for MZ and DZ twins, whereas the pathways for non-shared environmental variance and covariance across twins are set at 0 for MZ and DZ twins.

Fig. 3

The psychometric model for child temperament based on mothers’ and fathers’ reports. Measured variables are represented by triangles and latent variables by circles. The upper part of the model (above the measured variables) contains a common factor of child temperament which captures the shared variance between mothers and fathers. A1, C1, and E1 (where 1 stands for twin 1) are the genetic and environmental estimates for this common factor. Double-headed paths represent the correlations between the additive genetic variance (MZ = 1.0; DZ = .50) and shared environmental variance (MZ/DZ = 1.0). The lower part of the model (below the measured variables) contains the effects on the variance that is unique to each rater: unique genetic variance (Am1 for mothers’ report), unique shared environment (Cm1), and unique non-shared environment (Em1). The paths representing the factor loadings for each of the informants on the common factor were fixed to one for model identification.

Fig. 4

The Rater Bias model for child temperament according to mothers’ and fathers’ reports. The same convention as for the psychometric model is used in the Rater Bias model with the difference being in the lower part of the model. The set of unique estimates is composed of the rater bias estimates (e.g., bm where m stands for mothers’ report) and their corresponding residuals (e.g., M1). The single-headed paths (e.g., bm and bf) are the effects of the rater bias and the residuals estimates on the measured variables.