Global Information for Multidimensional Tests
Katherine Grace Jonas · 2019
New measures of test information quantify information that could be gained by administering the test to an examinee of unknown ability—termed global test information. Currently, these measures have been developed only for unidimensional tests. This study develops measures of multidimensional global information, validates them in data are simulated from models of varying dimensionality, and applies them in neuropsychological data from Rush University's Memory and Aging Project to identify the most informative tests of Alzheimer’s disease. These measures allow for direct comparison of complex tests normed in different samples, facilitating test development and selection.