Learning is the focus of attention for any educational enterprise. Consequently, the quality of assessment of the extent to which learning has occurred is of crucial importance. Assessment plays an important role in education in Hong Kong as in other places. Nevertheless, how to design an assessment that is of high quality, or how to fully utilize the results of the assessment to understand more deeply about the ability of our students is a demanding task for most teachers. It is not uncommon for teachers to assume that those students who cannot get good assessment results are weak students or that they have learning problems. Teachers rarely question the quality of assessment items to be the source of problems. In other words, teachers seldom consider the possibilities that the items may be too difficult, or unsuitable for the students' levels. In fact, it is worth investigating both the students' and items' performances for the assessments in order to understand more about the ability of our students. Under such circumstances, Rasch Measurement Model is a tool that teachers may consider to assist them for interpreting and analyzing the alignment between student ability and assessment item difficulty. In Rasch modelling, the data are checked, using a set of procedures and criteria, for the extent to which they match with the theoretical model. In particular, the point-measure correlations, and the INFIT and OUTFIT statistics are used to assess the goodness of fit between model and data. This study aims to illustrate how teachers can make use of these statistics from Rasch modelling to support them in the implementation of learning oriented assessment. The study will be based on assessment data obtained from two classes of Primary Three Students in a local school in July 2006. In the study, we will focus on the implication of misfit persons and items by using Rasch Measurement Model so as to discuss ways for understanding our students and improving the items and setting
