Validation of a Short Scale for Student Evaluation of Teaching Ratings in a Polytechnic Higher Education Institution

Abstract

The general purpose of this work is 2-fold, to validate scales and to present the methodological procedure to reduce these scales to validate a rating scale for the student evaluation of teaching in the context of a Polytechnic Higher Education Institution. We explored the relationship between the long and short versions of the scale; examine their invariance in relation to relevant variables such as gender. Data were obtained from a sample of 6,110 students enrolled in a polytechnic higher education institution, most of whom were male. Data analysis included descriptive analysis, intraclass correlation, exploratory structural equation modeling (ESEM), confirmatory factorial analysis, correlations between the short and long form corrected for the shared error variance, gender measurement invariance, reliability using congeneric correlated factors, and correlations with academic achievement for the class as unit with an analysis following a multisection design. Results showed four highly correlated factors that do not exclude a general factor, with an excellent fit to data; configural, metric, and scalar gender measurement invariance; high reliability for both the long and short scale and subscales; high short and long-form scale correlations; and moderate but significant correlations between the long and short versions of the scales with academic performance, with individual and aggregate data collected from classes or sections. To conclude, this work shows the possibility of developing student evaluation of teaching scales with a short form scale, which maintains the same high reliability and validity indexes as the longer scale.

Motivated Strategies for Learning Questionnaire (MSLQ): Adaptation, validation, and development of a short form in the Chinese context for mathematics

“…Third, in line with previous research (e.g., Gogol et al, 2014;Sánchez et al, 2021), our short form was administrated as part of the corresponding full scale. This may overestimate the relationship between short form and full scale because of the shared correlated measurement error.…”