Implementation of JMLE Procedure for the Rasch Model
Birnbaum's joint maximum likelihood estimation (JMLE) paradigm will be implemented in a BASIC program using the Rasch one-parameter logistic item characteristic curve model. This model was employed here in order to have a compact program illustrating the JMLE paradigm. Again, it is assumed that the initial data processing tasks that produce the vectors of number of examinees with each possible test score and number of examinees answering each item correctly have been performed beforehand. Since the item parameter and ability estimation stages are performed in an alternating manner until convergence is reached, the driver (kernel) section of the program is more involved than those of the preceding appendices.