The latent structure of spatial skill: A test of the 2 × 2 typology.

Multiple frameworks for categorizing spatial abilities exist but it has been difficult to verify them using exploratory factor analysis. The present study tested one of these frameworks-a 2 × 2 classification scheme that crossed the dimensions of static/dynamic and intrinsic/extrinsic (Uttal et al., 2013)-using confirmatory factor analysis with data on spatial performance from kindergarten (N = 251), third grade (N = 246) and sixth grade students (N = 241). For kindergarten and third grade students, four models were tested at each grade level: A 1-factor model, two 2-factor models (one static vs. dynamic, the other intrinsic vs. extrinsic), and a 4-factor model. In sixth grade, only the 2- and 1-factor models could be tested given the available data. Evidence that the 4-factor model was the best fit would have validated the 2x2 model. However, the 4-factor models failed to converge in kindergarten and third grade. Both the 1- and 2-factor models converged in these age groups, and chi-square tests demonstrated that the 2-factor intrinsic-extrinsic model was the best fit at both grade levels. In sixth grade, only one of the 2-factor models converged and it did not fit significantly better than the 1-factor model. Thus, there was limited validation of the model in these grades, as well as a trend toward less dimensionality in spatial skill over development.