Rasch Analysis on Reliability and Fit Statistics of a Self-Efficacy Measure With Different Rating Categories

• In Event: Rasch Measurement Innovations

Abstract

The Self-Efficacy Questionnaire for Online Learning (SeQoL; Tsai et al., 2020) was administered with a 0–4, 0–6, or 0–10 scale to 651participants, to compare the reliability and fit statistics across the different forms. Our preliminary findings suggested that across the three forms of the rating scale, there was little difference in item statistics, total scores, or Cronbach’s alpha. However, in terms of criterion validity, we observed that the SeQoL 11 (0-10) scale showed poor evidence of criteria validity. Second, we observed that the values of model error for item estimates increased as the number of categories increased. Finally, across the three forms of the rating scale, there was little difference in the reliability and separation measures.

Authors

• Chia-Lin Tsai, University of Northern Colorado
  Presenting Author

• Samantha Estrada, University of Texas - Tyler
  Presenting Author

• Julia Fulmore, University of Dallas
  Non-Presenting Author