This study aimed at investigating the dimensionality of the Intern Teaching Evaluation Form (ITEF), which is an observation instrument used in the evaluation of teaching practice at a public university in Ghana. The instrument was used to assess the teaching skills of a group of interns across three lessons. Data were obtained over a three-year period from 2016/2017 to 2018/2019 academic years during the Students Internship Programme (SIP) of the university. Part of the overall aim was to find the score stability (reliability) of the ITEF when it is used to evaluate teaching skill. A random effect two facets fully crossed design where intern (t) was crossed with item (i) and lesson (l), given by (t×i×l) was adopted for the dimensionality investigation. Further, a random effect one-facet fully crossed design, where inter (t) was crossed with lesson (l), given by (t×l) was adopted for the reliability investigation. The sample sizes for the study were 9132 undergraduate bachelor’s degree ITEF triplicate scores for the reliability investigation and 50 ITEF triplicate scores for the dimensionality investigation, selected by purposive and simple random sampling methods respectively. Univariate generalisability analysis using EduG was performed to analyse data. Findings were that, the ITEF is not unidimensional, but highly reliable (strongly stable) across the delivery of three lessons. Caution should be exercised in taking the ITEF scores as unidimensional in running correlational analyses with other criteria, unless this is done only from the perspective of ‘use,’ which should be one of the important considerations in instrument design.
Keywords: : Teaching practice, Reliability, dimensionality
