Abstract

This article presents two proposals for edumetric calibration coefficients for a test, specifically for quantification of validity, based on factorial analysis techniques and for quantifying the difficulty of a test or item. These proposals overcome some of the limitations of the coefficients proposed in classical test theory and item response theory. It's not consistent, difficult items without measuring what we want, that is, the difficulty should not be a stranger edumetric element and therefore must relate to reliability and validity of the test.