Speaker: Phil Chalmers, York University
Department of Psychology

Title: Generating Adaptive and Non-Adaptive Test Interfaces for Multidimensional Item Response Theory Applications

Abstract: Computerized adaptive testing (CAT) is a powerful technique for improving measurement precision and reducing the total number of items required in psychological tests. CATs are constructed by forming tailored test forms uniquely suited to each participant by capitalizing on the information available from responses to previous items. CAT applications primarily rely on the theory of unidimensional item response theory (IRT) to help select which items should be used during the computer session. However, multidimensional CAT can be constructed to further improve measurement precision and reduce the number of items required to obtain suitable measurement reliability for multiple psychological traits simultaneously.

This presentation describes a new R package for implementing unidimensional and multidimensional CATs for a wide variety of IRT models that can consist of dichotomous and polytomous responses, and demonstrates how powerful graphical user interfaces for unidimensional and multidimensional CAT applications can be generated using the mirtCAT package (Chalmers, submitted). Various examples of adaptive and non-adaptive interfaces (i.e., surveys) will be demonstrated, and the responses collected from these interfaces will be analyzed using tools from the mirtCAT package.

Suggested Readings:

Chalmers, R. P. (submitted). Generating Adaptive and Non-Adaptive Test Interfaces for Multidimensional Item Response Theory Applications. Journal of Statistical Software. PDF available by request only (email request to: rphilip.chalmers@gmail.com)

Magis D, Raiche G (2012). “Random Generation of Response Patterns under Computerized
Adaptive Testing with the R Package catR.” Journal of Statistical Software, 48(8), 1–31.
ISSN 1548-7660. URL http://www.jstatsoft.org/v48/i08.