Using the Rasch model to examine pre-service teachers approach to learning profiles

Abstract

Examining the manner in which learning is organized for educators is an important aspect of educational research. However, one main overarching methodological concern with previous studies was the consistent treatment of Likert scales as an interval type data, that may have resulted in slight inaccuracies in the results. This study thus aims to introduce an alternative method using a Rasch model that may be more suitable to understand the learning profiles of educators. Participants were 352 students (106 male, 246 female) enrolled in a postgraduate program in education. The theoretical framework was based on deep and surface learning approaches and was measured using the instrument RASI. The measurement tool was first validated using the Rasch Rating Scale model and the ability measures of deep learning and surface learning for each student were computed. Analysis indicated that the data was an excellent fit for the Rasch model and that the categories for all items were clearly ordered. Both of the subscales captured most of the spread of person measures. Subsequently, a person-orientated approach was used on the ability measures to categorize the pre-service teachers. The cluster analysis on these ability measures indicated that four groups of students differing in their levels of deep and surface approach to learning were obtained, namely D+/S−, D+/S+, D/S−, and D−/S. Comparisons across clusters for age and gender showed that students with a high deep and low surface approach were significantly older than others. In summary, the present study has demonstrated an alternative method of cluster analysis to identify different learning profiles, by transforming ordinal survey data to an interval level data.