Australian Research Council Linkage Grant LP0454080 Research Papers
Note for readers
In 2003, the Performance Measurement and Reporting Taskforce (PMRT) applied successfully as an industry partner with Professor David Andrich and Murdoch University for Australian Research Council (ARC) Linkage Grant LP0454080. The project was completed over a three year period.
The project was focused on large scale assessments which involve comparisons of the performance of students across time. For many reasons, identical items cannot be administered to every student on every occasion. Nevertheless, the items must operate invariantly across groups. The Rasch models, which are used in Australia for scaling items, provide a powerful means for checking for such invariance. This project brought together the basic research of Professor Andrich and his colleagues to study, apply and document ways in which violations of Rasch models and different data collection formats are reflected as changes of scale.
The research was relevant in the assessment of all areas of assessment including writing. The reason that the topic is significant for national assessment is that differences in scale can make the percentage of students reported as achieving the benchmark misleading. The project focused on how differences in scale might be controlled both statistically and empirically. In the case of writing, the research indicates that marking keys and formats need to be modified in order to control the effects on the scale rather simply statistically.
During the course of the project two workshops were convened for State and Territory psychometricians. The workshops covered relevant measurement theory, practical matters related to the construction of consistent scales, as well as a consideration of the equating issues.
These papers do not necessarily represent the views of either MCEETYA Ministers or PMRT members.
Eleven research papers were developed as outcomes of the grant project. The research papers may be downloaded in PDF format below.
ARC Linkage Grant LP0454080 Research Papers
Andrich, D. (2006). Understanding the response structure and process in the polytomous Rasch model.
Andrich, D. (2006). On the fractal dimension of a social measurement: I.
Andrich, D. (2006). On the fractal dimension of a social measurement: II.
Heldsinger, S. and Humphrey, S. (2006). Maintaining consistent metrics in standard setting.
Humphrey, S. (2006). The impact of differential discrimination on vertical equating.
Humphrey, S. (2006). Maintaining a common unit in social measurement.
Marais, I. (2006). RUMMss manual.
Marais, I. and Andrich, D. (2005). Studies on the effect of violations of local independence on scale in Rasch models: The dichotomous model.
Marais, I. and Andrich, D.(2006). Effects of varying magnitude and patters of response dependence in the unidimensional Rasch model.
Marais, I. and Andrich, D. (2007). Formalising dimension and response violations of local independence in the unidemsional Rasch mode
Van Wyke, J. and Andrich, D. (2006). A typology of polytomously scored mathematics items disclosed by the Rasch model: implications for constructing a continuum of achievment.