Item performance on the Stanford-Binet by a sample ( n = 289) of Georgia children (CA = 5) and by equivalent-aged children used in the test standardization was compared in terms of 5 subject variables: race, sex, socioeconomic status, intelligence level, and community size. The Georgia Sample's performance exceeded ( p < .01) the norm group's performance on 62% of the items. The five subject variables were associated with these differences: among levels of the variables, variations occurred in the presence and direction of item performance differences; generally, more than one subject variable influenced item performance. These data affirm that certain variables confound intelligence test performance and that norms based on a single variable, like race, do not eliminate test biases.
On the stability of the linear functional equation in a single variable on complete metric groups