ERIC Number: ED360368
Record Type: Non-Journal
Publication Date: 1993-Apr
Dimensionality of NAEP Instruments That Incorporate Polytomously-Scored Items.
Carlson, James E.
In this article some results are presented relating to the dimensionality of instruments containing polytomously scored as well as dichotomously scored items, concentrating on the 1992 National Assessment of Educational Progress' (NAEP) mathematics and reading assessment data and several simulated datasets. The maximum likelihood factor analytic procedure of the LISREL 7 computer program was used. Results were evaluated through use of residuals from the fitted model. The square root of the mean squared residual was the statistic used. Overall sample sizes for mathematics were 1,125, 1,173, and 1,064 for grades 4, 8, and 12, respectively. For reading, the sizes were 1,169, 1,271, and 1,139 for each grade, respectively. Results suggest that the dimensionality of data structures in the NAEP assessment is generally not affected by the inclusion of polytomously scored items, but the data structures cannot be generalized to other situations. One reason is the size of the correlations among the scales of the NAEP, and another is the small number of conditions simulated in this study. In addition, the number of polytomously scored items was limited in the 1992 assessment. Eight tables present analysis results, and four figures illustrate the square roots of the mean squared residuals (Contains 32 references.) (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: Educational Testing Service, Princeton, NJ.
Identifiers - Assessments and Surveys: National Assessment of Educational Progress