The Information Function Of Multidimensional Item Response Theory Model

The multidimensional item response theory and relevant models, the information function of unidimensional item response theory and the information function defined by Mark D.Reckase, the dispersion matrix derived from Daniel O.Segall and by which how we select items etc. are introduced in this paper. Also a new definition of the information function in MERT model based on the work of Daniel O.Segall is available by using dispersion matrix, Fisher information matrix.Based on the study of new defined information function, we can find:(1) Test information function is not equal to the cumulative sum of separate item information function under multidimensional situations.(2) It is possible to represent the test information function of component of ability parameter vector through measuring other variables.(3) In two dimensions situation, Item parameters, which include discrimination parameters α1, α2and intercept parameter d, and ability parameter θ2have the effect of changing the location where the maximum value of the item information function of θ1is attained.The surfaces of item information function are, in general, bell shaped, while α1affects the steep degree, α2affects the rotation and translation is influenced by parameterd.