Items Reduction in Course Evaluation Survey Questionnaire: An Analytical Appraisal

• Activity Number: 353
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Education
• Author(s): Abdullah Al Rubaish and Lade Wosornu and Sada Nand Dwivedi*+
• Companies: University of Dammam and University of Dammam and University of Dammam
• Keywords: Item reduction ; Course evaluation survey ; Factor analysis ; non-co-colinearity ; inter-correlation ; overall reliability

Abstract

The University of Dammam (UoD) is developing robust evidence required for academic accreditation. One of the important tools is course evaluation survey (CES) questionnaire. This study aims to employ theoretical, pragmatic and analytical considerations to optimize the number of items in this questionnaire while preserving its theoretical structure and analytical power. For exploratory analysis, the currently available CES data in four colleges (nursing, medicine, dentistry and engineering) were used. The college-specific number of courses ranged between 1 to 15 and the number of surveyed students from 32-113. Evidence of co-linearity among the items paved the way to systematically reduce the number of items in the original questionnaire. The analytical methods included consideration of data collected on ordinal scale as internal scale, correlation matrices and factor analysis. In spite of the modest sample size, three desirable characteristics of a questionnaire (non-co-linearity among items, inter-correlation among items, and overall reliability) showed either improvement or more consistency in the shortened questionnaire than that in original questionnaire. In summary, the origin