The Reliability of Multi-target Assessment Tryouts: A Case Study for Teaching and Learning Assessment at Bangkok University

Abstract

This study investigates the reliability of the multi-target assessment tryouts calculated from both main group and subgroups, and the relationships between the number of respondents and the reliability of the assessments. The subjects of this study are 57 Bangkok University students, categorized into 6 groups, with 5-15 students in each group The statistics used in this study are corrected item-total correlation, Cronbach’s alpha if item deleted, mean, standard deviation, reliability, Fisher’s Z, and Pearson’s correlation coefficient.

The results show that the reliability calculated from the subgroups is at the same level with that from the main groups. Moreover, the number of respondents and the reliability are not statistically significant at the level of 0.05. Therefore, to try out any assessment, the subjects may be recruited from students in any class taught by any number of instructors.