The Use of a Factor‐analytic Model for Assessing the Validity of Group Comparisons

This paper assesses the validity of a comparison of mean test scores for two groups of students, and of a longitudinal comparison of means within each group. Using LISREL, confirmatory factor analyses were carried out (a) to test the hypotheses of similar factor patterns, equal units of measurement, and equal accuracy of measurement between the two groups, and (b) to estimate the correlation between the latent traits measured by two successive test administrations in each group. The results indicate (a) that a comparison of the group means may be invalid because, although the factor pattern was the same for both groups, the factors were not measured in the same units, and (b) that longitudinal comparisons within each group are seriously complicated by evidence of structural change.