Abstract
Traditional methods of scaling to render school internal assessments comparable across schools require one and only one set of scores to play the role of scaling criterion. If more than one set of scores are available, each with varying degree of relevance to the assessments to be scaled, can we use them all to be scaling criteria? This paper introduces an objective method to put weights on the available scaling criteria and form a composite out of them. It is shown that this composite scaling criterion is better than any one of its components alone.
The method puts weights on the sealing criteria depending on their relevance to the assessment to be scaled. It utilizes multiple regression and works in an iterative process which starts from any initial set of weights input by the user. It then renders them to converge to a final set of weights which is independent of the initial set of weights. Concrete examples are given.
Keywords: | school internal assessments; scaling criteria |
---|
[Chinese Version | Index | New Horizons in Education | Other Journals | Subscription form | Enquiry ]