Application of Item Response Theory (IRT) to PISA

Organised by Hong Kong PISA Centre, HKIER & OECD/PISA Secretariat
This workshop on application of IRT to PISA will take place on 27-29 March 2006 in Hong Kong and will be conducted by Dr. Margaret Wu.
Objectives and rationale
The initiative for this workshop originated from requests by the PISA Governing Board to organise further training workshops for capacity building in the area of PISA data analysis. Scaling student performance in order to generate student proficiency scores is an essential area of PISA and requires advanced understanding of IRT, which is vital for any further advanced PISA data analysis. Therefore, the primary aim of the workshop is to give participants an overview of the methods and procedures used in PISA scaling and to familiarise participants with the application of IRT scaling approaches in the PISA context using ConQuest (1998) software.
ConQuest is an example of Item Response Modeling software that can fit a number of different item response models, including the simple Rasch model (Rasch, 1980), the partial credit model (Masters, 1982), the many-facet model (Linacre, 1994), the multidimensional item response model (Adams, Wilson and Wang, 1997) and the latent regression model (Adams, Wilson and Wu, 1997). The scaling of PISA data was carried out using ConQuest, including the calibration of item parameters and the generation of student proficiency scores.