Educational Researcher, Vol. XX No. X, pp. 1–11
DOI: 10.3102/0013189X20933836
Article reuse guidelines: sagepub.com/journals-permissions
© 2020 AERA. http://er.aera.net
MONTH XXXX 1
C
lass size is an issue that continues to stimulate debate
among educators, researchers, administrators, and poli-
cymakers because of its link to student achievement, the
quality of teaching and learning, and program evaluation.
Furthermore, class size has also received considerable attention
from stakeholders over the past quarter century given its eco-
nomic implications, particularly for institutions of higher educa-
tion (IHEs) (Krueger, 2003). In a 2014 report published by The
Chronicle of Higher Education that examined state aid to public
colleges between 1987 and 2012, 164 out of 169 research uni-
versities reported a decline in state funding within that period,
with nearly one in three reporting a decrease of 25% or more.
Consequently, IHEs may seek to increase class size as a means of
reducing human resource costs.
Most studies related to class size, whether K–12 or higher
education, tend to be highly quantitative (e.g., Diette & Raghav,
2015; Hanushek, 1986; Hoxby, 2000; Kokkelenberg et al.,
2008). However, their conclusions regarding the influence of
class size on student achievement are equivocal at best. In the
K–12 literature, some research has suggested that class size does
influence student achievement at various points in the K–12
experience (Borland et al., 2005). However, other research has
suggested that teacher efficacy has a more salient influence on
student achievement (Hoxby, 2000; Rockoff, 2004). When
studying IHEs, researchers have similar disagreements as to the
influence of class size on student achievement and its implica-
tions (Keil & Partell, 1997; Kokkelenberg et al., 2008; Terenzini
& Pascarella, 1991; Williams et al., 1985). Methodologically,
these class size studies employ a diverse array of models that
include student achievement controls and classroom environ-
ment variables; however, very few control for instructor experi-
ence or have attempted to parse the effect of class size by
simultaneously including race, gender, and academic discipline
within the same model.
Given the methodological limitations of previous higher edu-
cation class size studies, there is a need for a more robust quanti-
tative analysis that incorporates a broader range of student and
class-level variables, including those that control for instructor
experience and student experience. The goal of this study is to
present such a comprehensive model aimed at addressing this
gap in the literature.
933836EDR XX X 10.3102/0013189X20933836Educational ResearcherEducational Researcher
research-article 2020
1
Temple University, Philadelphia, PA
2
University of South Florida, Tampa, FL
Does Class Size Matter in the University Setting?
Ethan Ake-Little
1
, Nathaniel von der Embse
2
, and Dana Dawson
1
University class size is a frequently debated topic among stakeholders given its relation to student achievement, teaching
and learning, program evaluation, and education economics. However, the extant literature in both K–12 and higher
education contexts regarding class size is equivocal, with some citing evidence of an adverse effect on student achievement
for larger class sizes and others suggesting minimal effect. This study aims to explore the relationship between class size
and student achievement, as measured by course grades, in the core undergraduate program at Temple University, a
large, state-related university in Philadelphia, Pennsylvania. A cross-classified multilevel model was employed consisting
of 14 variables—6 student level and 8 class-level—and drawing from a robust sample size of 172,516 grades awarded to
32,766 students in 8,049 classes offered across 14 terms. Results suggest that, after controlling for instructor experience,
the effect of class size is not uniform and is, in fact, quite variable when accounting for student race, gender, and academic
discipline within the same model. We discuss the possible reasons for these variable results with implications for program
policy and classroom practice. Finally, we discuss the limitations of the present study and how future research might resolve
those limitations.
Keywords: class size; multilevel modeling; program evaluation; student achievement; university teaching and learning
FEATURE ARTICLES