Home

The mcIRT package was developed to provide useful functions for practitioners who have to analyze responses from multiple choice items, which are somewhat difficult to deal with. We know that item difficulty relies strongly on the choice of distractors, and we know that perhaps different options attract groups of persons in a different way, which could lead to differential item functioning (DIF). In fact, dichotomous item response models are considerably more often used than models which regard all the options. This will hopefully change step by step with every new release of this package. It includes two powerful models to evaluate multiple choice test data or any data which stems from any other scenario in which one option is chosen over some other options. There are a lot of options and finetunings (and hopefully they are getting more and more) that can be done with this models. So: Introduction End.

There are several topics I want to write about:

0. Some small simulations for experiencing estimation time and accuracy of parameter estimation
1. Simple model fitting procedures
2. Testing models against each other
3. Fixing item parameters
4. Estimating models with nonparametric ability distribution
5. Manipulating the design matrix
6. Creating plots with the results.