Abstract
[A Comparative Study of Scoring Criteria Used in Multiple Classification Tasks: A Cross-Sectional and A Longitudinal Study]
The relative merits of three types of scoring criteria for multiple classification tasks the Genevan judgment-plus-explanation criterion, the counter-suggestion criterion, and the non-Genevan judgment-only criterion — have been evaluated in two studies. The 487 children in the cross-sectional study, aged 4 to 11, completed ten multiple classification tasks at a single test administration. The 275 children in the longitudinal study, aged 4 to 7, were administered the same battery of tasks three times with an interval of four months. Scored in accordance with each of the three criteria the battery proves an internally consistent test for multiple classification. In each case, factor analysis yields a single factor and the factor structures show a very high degree of similarity. Furthermore, these factor structures remain stable in the longitudinal study, so that the multiple classification tasks refer to one and the same underlying concept for all three scoring criteria. Both cross-sectional and longitudinal findings indicate that major increments in multiple classification scores occur in the primary school grades 1 and 2, whereas development in preschoolers is minimal. Implications of these findings are discussed. The point is made that, in spite of the excellent psychometric properties of all three criteria, the selection of scoring criteria in future multiple classification research still has to be based on a firm theoretical framework.