![Sample our Education journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5932/TOC-Subject-Sample-2021-Education.jpg"})
Abstract
Structural equation modelling is used to measure the existence of a response style (in particular, acquiescence) behind three balanced Likert scales measuring different concepts in a questionnaire for student evaluation of teaching in higher education. Exploration with one sample (n = 1125) and confirmation in a second sample (n = 710) from a different population (different students, different courses) shows the existence of a common factor behind these scales. However, the weak correlation with the variable ‘sum of agreements’ does not support the idea of acquiescence in student evaluation of teaching. Instead, the common factor could be interpreted as a halo factor influencing different scales in the questionnaire. The implications of this exploratory study for future research on acquiescence in student evaluations of teaching are discussed.
Notes
1. Flanders is the Flemish‐speaking part of Belgium.
2. Billiet and Davidov (Citation2008), however, point out that this approach holds the assumption that positively and negatively worded items are equally affected by acquiescence and that the scales’ reliability indices (e.g. Cronbach’s alpha) will be rather low (which is, of course, better than artificially high alpha scores on scales with only positively worded items due to acquiescence).
3. The item loadings in an unscaled three‐factor model vary between .63 and .94. The negatively worded items of course have negative loadings.
4. Eleven positive answers: 5.9%; 12 positive answers: 2.0%; 13 positive answers: 0.3%; 14 positive answers: 0.3%.
