MODEL RESPON BERTINGKAT UNTUK KELAS LATEN MULTIDIMENSIONAL TEORI RESPON BUTIR (GRADED RESPONSE MODEL FOR MULTIDIMENSIONAL LATENT CLASS OF ITEM RESPONSE THEORY) Studi Kasus: Analisis Validitas Perangkat Tes Hospital Anxiety and Depression Scale

Adzan, Achmad Yusuf A, Danardono

2015 | Skripsi | FMIPA UGM

Item Response Theory (IRT) models are commonly used to measure latent traits, through the analysis of data deriving from the administration of questionaires made of items with dichotomous (binary) or polytomous. Traditional IRT models are based on the assumption of unidimensionality, which means that all items contribute to measure the same latent trait. Moreover in some cases, the assumptions of normality of this latent trait is explicitly introduced, unfortunately in several practical situations both assumptions are restrictive. The models at issue extend traditional IRT models allowing for multidimensionality and discreteness of latent traits. They also allow for different parameterizations of the conditional distribution of the response variables given the latent traits, depending on both the type of link function and constraints imposed on the discriminating and difficulty item parameters. These models may be estimated by maximum likelihood via an Expectation-Maximization algorithm. In this mini thesis discussed the analysis of the validity of the tests Hospital Anxiety and Depression Scale (Zigmond and Snaith, 1983) which consists of fourteen items using graded response model for multidimensional latent class of item response theory. The questionnaire is expected to be able to distinguish two different latent trait, namely the level of anxiety and depression. Having analyzed the above methods lead to the conclusion rejection bidimensionality assumptions latent nature so that all items in the questionnaire measuring the same latent trait, namely psychological disorders. Then on the fourteen items, item 2, 4, 6, 7, 8, 11, 13, 14 are eliminated from the questionnaire because it was not good enough in providing information at the level of certain psychological disorders.

Kata Kunci : item response theory, polytomous response, multidimensionality, discreteness, latent traits, maximum likelihood estimation, Expectation-Maximization algorithm.