Sibling imitation and contrast effects

A general linear model is developed to account for the effects of direct phenotypic imitation and contrast of sibs on one another. Specific predictions from an imitation/contrast model depend upon three assumptions. They are (1) the type of process (imitation or contrast); (2) the stage of the process at the time phenotypes are measured; and (3) sibling constellation variables such as the number, sex composition, and age distribution of a sibling pedigree. For some models, phenotypic variance becomes a polynomial function of the imitation/contrast parameters and the covariances between the genotypes and the environments of sibling pairs. Consequently, phenotypic variances of different types of siblings [e.g., monozygotic (MZ) twins versus dizygotic (DZ) twins versus foster sibs] provide information about imitation or contrast. Thus, standardization of measures prior to analysis may unwittingly hide the statistical information that could detect these effects. The types of data needed to analyze imitation and contrast effects and the potential power of resolving these components of variation are discussed.This work was supported in part by Grants GM-30250, GM-32732, and AG-04954.