A study of ordering mechanisms for generating R6 and SAI distributions from criteria-based assessment data

This study explores the issue of generating ordered distributions of Queensland senior secondary school students using criteria-based assessment data. Criteria-based assessment data are multidimensional, yet ordered distributions are unidimensional since order is a property of univariate data. Ordering multivariate data requires the data to be transformed onto a univariate scale and such transformations are called here ordering mechanisms.

The ordered distributions obtained from the criteria-based assessment data are used for certification purposes and tertiary entrance selection purposes. Importance is placed on these distributions, yet there seems to be little study of the ordering mechanisms to yield these distributions from multidimensional criteria-based assessment data. This study seeks to address this gap.

The purpose of this research is to explore the issue of generating macro-function assessment data from micro-function assessment data. More specifically, this research considers the issue of ordering multivariate educational assessment data for percentage-based assessment systems, and in particular, the nature of the ordering mechanisms for ordering systems within such assessment systems.

This study is focused by the following research question:
RESEARCH QUESTION:
Is there a mathematical system which: - orders all points in a multidimensional percentage-based assessment space,
- models preferences within a multidimensional percentage-based assessment space,
- can be used to construct an independent ordering system for a percentage-based assessment system?

The existence of a mathematical system for modelling ordering mechanisms has been questioned. This study shows that a mathematical system with properties which suggest its applicability for developing ordering mechanisms in assessment systems for certain Board subjects, does exist.

If you are the author of this work and have concerns regarding the content. Please contact Library [mailto:LibResearch@acu.edu.au] Research Services.