Publisher's Synopsis
The TESTFACT program performs item analysis and test scoring according to the principles of classical test theory. Designed originally for a national testing service, the program has all of the features needed for processing data from binary scored tests, subtests, or scales. Classical item statistics are computed, including measures of reliability, generalizability, and validity against one or more external criteria. The tests, subtests, or scales are scored, and the scores are summarized for designated groups of respondents.
As an aid to item selection and interpretation of scales, the program provides highly developed and rigorous procedures of item factor analysis. Capable of performing classical item analysis on as many as 500 items and unlimited numbers of respondents, the program can analyze up to 15 response alternatives including Omit and Not Presented codes. The data can consist of multiple subtests and multiple classes of respondents; detailed class by subtest results are displayed.
