Objective measures of shyness in the MacArthur Longitudinal Twin Study were obtained in two testing situations: the laboratory and the home. A longitudinal hierarchical model was fitted to the data, allowing estimation of the extent to which genetic, shared environmental, and unique environmental influences contributed to continuity and change of the shyness phenotype from 14 to 20 months of age. The sample consisted of 163 monozygotic and 138 same-sex dizygotic twin pairs. Models were fitted to raw data using a maximum-likelihood pedigree approach. Genetic, shared environmental, and unique environmental first-order factors, with specific variances, were modeled on each of four shyness ratings assessed in the laboratory and home at 14 and 20 months. Four second-order genetic, shared environmental, and unique environmental factors were also modeled. Results indicated that developmental change from 14 to 20 months and situational specificity between the laboratory and the home are mediated largely by shared and unique environmental influences. Genetic variation is largely responsible for both the stability in shyness from 14 to 20 months and the phenotypic correlations observed between the laboratory and the home settings.