A multilevel version of the latent Markov Rasch Model is proposed for the analysis of binary longitudinal data with covariates concerning pupils'proficiency. The model is formulated si that the distribution of the latent process, representing the evolution of the ability of each student, depends on random parameters related to the cluster to which the student belongs. Then, the effect of each cluster may be evaluated according to specific characteristics such as the type of school. Likelihood inference for the model, based on the EM algorithm, is outlined. An application based on data collected in an Italian Region is described
Bartolucci, F., Pennoni, F., Vittadini, G. (2010). Assessment of school performance through a multilevel latent Markov Rasch model. In Proceedings of the 25th International Workshop on Statistical Modelling July 5-9, 2010 Glasgow (pp.79-84).
Assessment of school performance through a multilevel latent Markov Rasch model
PENNONI, FULVIA;VITTADINI, GIORGIO
2010
Abstract
A multilevel version of the latent Markov Rasch Model is proposed for the analysis of binary longitudinal data with covariates concerning pupils'proficiency. The model is formulated si that the distribution of the latent process, representing the evolution of the ability of each student, depends on random parameters related to the cluster to which the student belongs. Then, the effect of each cluster may be evaluated according to specific characteristics such as the type of school. Likelihood inference for the model, based on the EM algorithm, is outlined. An application based on data collected in an Italian Region is described
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
