IACAT provides pre-conference workshops as a learning opportunity for attendees. Below are the workshops currently planned for 2019.
You will have the option of selecting one of the workshops when registering for the conference. Click here to register!
Introduction to IRT and CAT
John Barnard (EPEC) and David Weiss (University of Minnesota; Assessment Systems)
This workshop provides a broad overview of item response theory (IRT) and computerized adaptive testing (CAT) for those who are newer to the field. We assume a knowledge of basic psychometrics such as classical test theory. The workshop begins with a background on (IRT), how it can be used to evaluate item and test performance, and how it provides a number of improvements over the classical approach. We then provide an introduction to CAT, describing the components and algorithms necessary to build an effective CAT program: item bank calibrated with IRT, starting point, item selection rule, scoring method, and termination criterion. Finally, we discuss important aspects regarding how you might evaluate and implement CAT for your organization.
Simulations and CAT
Theo Eggen (CITO) and Angela Verschoor (CITO)
This workshop explores the role that simulation studies play in CAT research and publication of real CAT assessments. Simulations are run by software programs that typically apply a CAT algorithm to a data set of fake (monte carlo) examinees, or real data from past assessments. This provides the researcher control over important algorithms such as items selection methods, exposure controls, and termination criterion, and therefore allows them to imagine and explore experimental designs to investigate aspects of their performance. This is also essential in the establishment of validity evidence for the publication of a real CAT.
Chun Wang (University of Washington)
Multidimensional IRT (MIRT) provides a strong innovation in measurement theory, which can translate to a new paradigm for CAT. Using MIRT can create a number of advantages, including making the CAT algorithm even more efficient, especially if the purpose of the assessment is to measure anything more than a simple, clear construct.