TEST THEORIES: CLASSICAL THEORY AND ITEM RESPONSE THEORY

For a correct interpretation and proper use of the psychometric properties of tests it is necessary to go beyond the mere empirical calculation, and know the grounds on which these calculations are based. To contribute to this understanding beyond the superficial handling of the psychometric formulas, the main goal of this work is to present, in a not technical way, the two most important theories that guide the development and analysis of most tests: Classical Test Theory and Item Response Theory. First, a historic note about tests and testing is made, indicating the evolution of tests according to the technical and statistical advances. The importance of test theories in order to develop and analyse tests is pointed out, and Classical Test Theory, including Generalizability Theory and Criterion Referenced Tests, is presented. After underlining the limitations of the Classical Test Theory approach, Item Response Theory is presented. Within this new framework some of the limitations of the Classical Test Theory find a proper solution. Finally both approaches are compared, emphasizing the importance of test theories for a correct use and interpretation of psychometric properties of the tests.