Do students’ evaluations of teaching and academic achievements correlate? A meta-analysis

Yükseköğretim Kurumları Destekli Proje, BAP Araştırma Projesi, 2025 - 2025

For decades, universities have relied on Student Evaluation of Teaching (SET) surveys to assess instructors’ teaching

practices, with the underlying assumption that highly-rated instructors enhance student learning more. These evaluations are

typically conducted in the final weeks of semesters, prior to grade announcements, and prompt students to rate various aspects

of their instructors’ teaching, such as knowledge, clarity, organization, accessibility, and fairness. SET scores serve as significant

indicators in decisions regarding faculty recruitment, promotion, tenure, and awards, impacting academic careers and institutional

recognition (Spooren et al., 2013; Stark & Freishtat, 2014; Uttl et al., 2019).

The varying outcomes in related research underscore the importance of consolidating extant findings to better

understand the relationship between SET scores and student achievements. According to the assumption, students’ SET scores

should correlate highly with their course grades. Therefore, a robust methodology like meta-analysis, which effectively

synthesizes results and minimizes bias, is essential for determining an overall effect size across numerous studies (Borenstein et

al., 2009). By applying meta-analysis, this study aims to offer a systematic perspective on the relationship between SET scores

and student achievements. Given the inconsistent findings in the literature (Benton & Cashin, 2012; Clayson & Haley, 2011;

Spooren et al., 2013; Stark & Freishtat, 2014; Uttl et al., 2017), this meta-analysis aims to examine research results by

thoroughly reviewing studies retrieved from the Web of Science (WOS), Scopus, and Google Scholar databases, which were

selected due to their inclusion of multidisciplinary studies. The review covers all publications available up to April 2025, without

any restrictions on the start date. The study aims to provide significant contribution by offering an updated synthesis of the

literature and incorporating moderator analyses to investigate how various factors might influence the level of correlations

between SET scores and student achievements (Chen et al., 2018).

Many existing studies employ a traditional random-effects model that assumes independence among effect sizes across

studies. However, studies often report multiple effect sizes for various variables (such as different learning outcomes) orfor

distinct sections within a single study. This approach does not adequately address nested data structures, leading to insufficient

handling of clustered effect sizes. To manage this dependency, multilevel meta-analysis will be applied, allowing for an explicit

modeling of dependencies between effect sizes at different levels within studies or subgroups, thus providing more precise

estimates of the true effect size (Van den Noortgate, 2014).

In this project, analyses that account for dependency through multilevel meta-analysis will be compared with those that

do not, offering insights into the impact of modeling choices on the overall conclusions. By comparing results from both

dependent and independent analyses, the study aims to underscore the importance of addressing effect size dependencies in

meta-analytical research, enhancing the accuracy of findings on the relationship between SET scores and student achievements.