Parameter Estimation of KST-IRT Model under Local Dependence

A mantra often repeated in the introductory material to psychometrics and Item Response Theory (IRT) is that a Rasch model probabilistic version of Guttman scale. The idea comes from observation sigmoidal item response function provides characteristic models an It appears, however, more difficult reconcile assumption local independence, which traditionally accompanies model, with dependence existing In recent work, alternative scale was proposed, combining Knowledge Space (KST) IRT modeling, here referred as KST-IRT. present work has, therefore, two-fold aim. Firstly, estimation parameters involved KST-IRT discussed. More detail, two methods based on Expectation Maximization (EM) procedure are suggested, i.e., Marginal Maximum Likelihood (MML) Gibbs sampling, compared basis simulation studies. Secondly, for scale, estimates those traditional combination plus independence under interchange data generation processes. Results show approach might be effective capturing it appears robust misspecification process, but price increased number parameters.