Duration: 4 hours
Description: Many large-scale assessment programmes around the world use item response theory (IRT) for test performance management.
This approach is often more than a statistical analysis method - it is a complete approach for designing tests and understanding how they work with candidates. It is particularly valuable in situations where there are various versions of tests that need to have results interpreted on a common scale (for example parallel forms, or tests used from year to year).
This session will give attendees an introduction to several such statistics; explaining their bases in test theory and giving practical illustrations of how they are used.
In this session, we will look at why IRT is necessary, the scenarios where it brings value, and how it can be used to calibrate items and tests. The session will also investigate IRT’s limitations, and how professionals can use it in a defensible manner.
- the item response function
- IRT models
- analysis of model fit
- a comparison of classical and item response theories
Typical Task – practical use of IRT for your assessments.