The problem of measuring attitudes in preschool children is discussed. The general rationale employed is closely related to that of the semantic differential (SD). Research employing the SD with older children and adults has demonstrated that the primary dimension of affective meaning is that of evaluation. It has also been shown that evaluation scores from the SD are closely related to scores obtained from conventional attitude scales. It was demonstrated that a comparable evaluation dimension is present at the preschool level, and procedures were developed for assessing attitudes in preschool children employing this rationale. Procedures are picture-story techniques in which the child is shown a picture containing two figures and is told a story in which a positive or negative adjective is employed. The child is asked to select the figure which he thinks is the one described in the story. This method has been used to assess preschooler's attitudes toward the colors white and black and toward human figures with light and dark skin-color. Findings and results of this and other studies are described and a new revised version of the Preschool Racial Attitude Measure is presented. Conclusions include: (1) A child's racial attitudes may be associated with certain parental variables, and (2) Racial attitudes may partially result from the child's early learning experiences involving light and darkness. (Author/CK)