the classical and modern theories of measurement in behavioral science and medicine: a review on the methodological properties, benefits and limitations

the classical test theory (ctt) and item response theory (irt) are widely used in construction and analysis of exams and questionnaires. because of the various shortcomings of ctt, irt has been designed. even though irt has overcome most of the limitations of ctt and has made computer adaptive testing possible, it is still important to point out that not enough reasons exist to totally discard ctt. this theory is still in use in different scientific fields. previous studies have showed that irt is complementary to ctt, and is not a competitive model. the main difference between these two theories is that ctt is test-based, but irt is item-based; for example, measurement error modeling in ctt and irt were done in the test and item level, respectively. in practice, the irt is shown to have very good performance in dealing with the problems presented in the ctt model; such as conditional standard errors, developing parallel tests, equating, and item bias and adaptive testing. in this study irt and ctt are explored and the weaknesses and advantages of both theories have been discussed.