This is a progress report on research conducted (1) to consider the methodological issues of response set and multidimensionality, which might lead to a refinement of the Test Anxiety Scale for Children (TASC) and the Defensiveness Scale for Children and (2) to investigate social background and school achievement correlates of test anxiety and defensiveness in a more heterogeneous group than had previously been done. The scales were expanded so that 7,551 second graders were randomly assigned to one of six possible test situations . Factor analysis of the original TASC for each sex revealed four factors, three of which were replicated on a factor analysis of the expanded TASC. None of the factors were interpreted as clearly defining response set. The four factors were used as subscales in a two-way multivariate analysis to test the race and sex effect. All four subscales contributed to the race effect, while the overall sex effect was due to the white sample only. When results were compared to total TASC scores, the effects proved significant. The following conclusions and directions were made: (1) acquiescent response set is not a major source of variance in TASC, (2) multidimensionality of the scale orients researchers to focus on childrens' anxiety responses to school situations, and (3) it would be difficult to use the interdependent subscales as differentiating correlates of TASC. (JS)