Comparing different item thresholds for Rasch rating scale model / Sholeh Ataei

Measurement issues are old issues in educational testing, and some of them are still unsolved. The Rasch rating scales model is a unidimensional measurement model for analyzing data on human attitudes. This mathematical model describes the two parameters: items and persons as agents of response. This study investigates the measurement of rating responses (attitude items) for the Rasch rating scales using item threshold models. If measure is merely based on the graphical distribution of the Wright map without considering the item thresholds, the item measure does not reflect the difficulty of achieving a particular category in the rating scale. Although previous studies have described and interpreted the rating scale category functioning and Wright map for the rating scales model, only few have described the importance of thresholds in calibrating the measurement variables of the rating scales. The mathematical and practical concepts of the measurement must be examined further to address these issues. Accordingly, this study investigates and estimates the consistency of the rating scale model categories and the model calibration between the parameters of measurement/latent variables in a particular category based on the Rasch mathematical functions and real data. The data are gathered by conducting a questionnaire survey among 342 engineering students taking up Calculus III at the Universiti Teknologi MARA in Shah Alam