| Juan Moran
University of Illinois
Examining depth and dimensionality of achievement: Application of the Unified Cognitive-Psychometric Diagnosis Model to the TIMSS 4th grade science data
Current advances in psychometric theory are Item Response Theory (IRT) based models which include parameters to represent knowledge structures and states reflecting the underlying cognitive processes of test performance. One of this models is the Unified Cognitive-Psychometric Model (DiBello & Stout, 1995, 1997). The model classifies students on the basis of their knowledge states as well as measures of their overall ability level. This classification (conceptual competency space) is defined by two dimensions, the ability level (theta from IRT) and a fit index (zetha) which measures the typicality of the students' response pattern. The Unified Model will be applied to achievement data from the 4th grade science TIMSS data in the content area of physical science using the different items types. Achievement comparisons within and between three countries (South Korea, U.S. and Portugal) will be done. Preparatory analysis will be done to assure degree or homogeneity in the samples and to correct statistical distributions due to sampling. Three types of dimensionality analysis will be done: differential item functioning, conceptual complexity and of the psychometric model. The Unified Model will be applied to look at the depth and dimensionality of knowledge of the science items. The various knowledge progress paths will be delineated from the conceptual competency space. This project will address policy issues related to the linkage of assessment, curriculum and the education. Influences of culture, language and gender will also be examined. It will also attempt to link the TIMSS 4th grade data to the NAEP data using the new psychometric models.
Back to Funded Dissertation Grants Page |