BAYESIAN ANALYSIS IN A SPEEDED ITEM RESPONSE MODEL J. BAZAN, L. VALDIVIESO * , M. D. BRANCO ** 1 * Departamento de Ciencias, Pontificia Universidad Cat´olica de Lima, Per´ u ** Departamento de Estat´ ıstica, Universidade de S˜ ao Paulo, S˜ ao Paulo, Brazil April 30, 2010 Abstract We develop a Bayesian analysis of the model proposed by Goegebeur, de Boeck, Wollack and Cohen (2008) to work with speededness by considering a gradual process change IRT model. Prior specification and MCMC algorithm are discussed. Simulations studies are conducted to evaluate his performance to the recover the parameters by considering independence or dependence in examinee´s parameter or personal latent variables by using independent priors. A real-data example, EGRA test, is analyzed and show how the model can be used to identified personal latent variables associated with response in Speededness situations which is consistent with the results in the test. Keywords: Item response models, Bayesian estimation, Speededness, guessing parameter, logistic model, DIC, simulation, EGRA . 1 Address for correspondence: Jorge Luis Bazan. Av Universitaria 1801, San Miguel, Lima 32, PUCP, Peru. E-mail: jlbazan@pucp.edu.pe. 1