In October, during the final 2012 U.S. presidential debate, the topic of class size came up within the context of global competitiveness. Although the candidates were mainly arguing the benefits of small classes in K-12 education, the issue deserves attention within higher education. With the growth of online classes, including massive open online courses (MOOCs), and with the creep upward in class size of many institutions that have faced budget constraints in recent years, it is worth asking whether class size matters in college courses. Do the learning objectives, teaching methods, teacher standards, and workload expectations vary, depending upon class size? Do students’ learning, motivation, and work habits in large classes match those in smaller classes?
Historically, research findings on the effects of class size on student achievement in postsecondary education are equivocal. Some have found no significant impact;  others have discovered that student achievement declines slightly as class size increases.  But such studies predate the widespread development of online classes with very large enrollments. More recently, others found that students in smaller classes reported learning more.  Medium-sized classes have been found to have little to no benefit over large classes with respect to undergraduate student learning and achievement in mathematics; however, student engagement is stronger in small classes.  With respect to distance education, class size may matter less  than students’ reasons for taking the course and their class status (i.e., lower-level vs. upper-level students).
The IDEA Center, a nonprofit organization whose mission is to serve colleges and universities committed to improving learning, teaching, and leadership performance, retains an archived database of student ratings that provides the opportunity to explore the effects of class size on faculty and student perceptions of learning and instruction. Student ratings of instruction collected from 2002-11 include undergraduate and graduate classes from public and private colleges across all regions of the continental United States. Approximately 40 percent come from master’s level institutions, 20 percent each from bachelor’s and doctoral level, 5 percent from associate, and 5 percent other. The data include ratings collected online and on paper, in face-to-face and online classes.
In our ratings system, the instructor indicates the course enrollment.  Instructors also rate the relevance of 12 learning objectives (minor or no importance, important, essential) for the course. In addition, they have the option of answering questions about their approach to instruction, course requirements, and various course circumstances. On the student ratings form,  students rate their progress on the same 12 objectives, the frequency of 20 teaching methods, and various course, teacher, and student characteristics. They also provide an overall rating of the course, instructor, and their attitudes toward the field of study.
Historically, the IDEA Center has categorized class size as small (10-14), medium (15-34), large (35-49), and very large (50+) when preparing technical reports. This same grouping was applied to the current analyses, which resulted in the following distribution of classes: small (63,622), medium (349,313), large (48,916), and very large (27,503).
The first thing evident is that the objectives instructors choose to emphasize vary by size of class. Instructors in very large classes are more likely to emphasize learning factual knowledge and less likely to stress developing communication skills (both oral and written) than are those in small and medium classes. This is especially true in general education courses.
As one might expect, the primary approach to instruction varies as well. One of the reasons instructors in large and very large classes emphasize the learning of factual knowledge may be because they rely upon lecture as the primary approach to instruction. Instructors in very large classes (about 86 percent) are more likely to lecture than those in small (43 percent) and medium-size (54 percent) classes. Still, lecture remains the most frequent teaching method regardless of class size.
Teaching methods differ as well. According to students, instructors in small and medium classes are more likely to involve students in hands-on projects and real-life activities, assign projects that require original or creative thinking, form teams or discussion groups to facilitate learning, and ask students to help each other understand concepts or ideas. Perhaps most troubling is that students in large and very large classes report the instructor is less likely to inspire them to set and achieve goals that really challenge them.
The reason for such lack of inspiration and challenge may relate to differences in course characteristics. Students in very large classes report fewer non-reading assignments than do those in small and medium-size classes. They also rate instructors lower on their achievement standards and their expectations that students share in responsibility for learning. So, the case could be made that students perceive larger classes as less rigorous.
But don’t assume students are champing at the bit to enroll in courses they perceive as less rigorous. Students in small classes consistently report a stronger desire to take the course than those in very large classes. Moreover, they report stronger work habits. And it’s not just the students who perceive such class-size differences. Fifty-three percent of instructors in small classes believe the level of student enthusiasm had a positive impact on learning compared to only 38 percent of instructors in very large classes.
Such enthusiasm translates into higher student ratings of progress on relevant objectives. Student average progress on course objectives the instructor rates as either essential or important is more than one-half standard deviation higher in small compared to very large classes. The advantage for small classes is especially evident in developing creative capacities (writing, inventing, designing, performing in art, music, drama, etc.) and communication skills (oral and written), where student progress is about a full standard deviation higher compared to very large classes. For medium-size classes, the advantage is nearly the same. When you compare small and medium-size classes with classes enrolling 100 or more students (of which there are over 6,000 in the database), the differences are even more staggering.
The smallest gaps in student progress between small, medium, and very large classes are found in gaining factual knowledge and learning fundamental principles and theories. The gaps do not even increase markedly in classes exceeding 100.
A finding particularly relevant for general education, where students sometimes get their first impressions of a discipline, is the relationship between class size and student attitudes toward the field of study. Students in small and medium classes report more positive attitudes about the discipline as a result of taking the course than do those in very large classes.
These effects of class size are not terribly surprising. The IDEA Center has known for years that class size makes a difference, which is why course enrollment has long been one of the variables we use to adjust student ratings scores. Moreover, recommended actions presented to instructors in the individual IDEA class report are made based on comparisons between the class’s average rating for a teaching method and other classes of similar size. The effectiveness of a teaching method depends not only on which objective is being emphasized but also on how many students are enrolled in the course. However, as reported previously and confirmed in the current dataset, student work habits and motivation are more important predictors of achievement on relevant learning objectives than is class size. The key is for faculty to encourage such productive behaviors in students regardless of how many are enrolled in the class. But in large and very large classes this is apparently a more daunting task.
Even in higher education, then, class size makes a difference. In very large classes, instructors are more likely to emphasize factual knowledge and less likely to develop communication skills. In turn, in very large classes students are less likely to report progress on communication skills and creative capacities, such as writing, inventing, designing, and performing. The types of learning where students in very large classes approach the progress of those in small and medium classes is in developing basic background in the subject matter.
As policy makers and institutions of higher education continue to explore the possibility of offering fewer sections with larger enrollments (including MOOCs and many other forms of online education and in-person education with large enrollments), the effects of class size on teacher behaviors and student learning, motivation, and work habits should be part of the conversation. Admittedly, the increasing sophistication of learning analytics and data mining has the potential of making MOOCs and very large classes more personalized. The instructor could have the ability to detect when a student is struggling and to provide targeted feedback and additional assignments to foster improvement, something that was previously likely only in small-to-medium size classes. Whether such an approach will support the development of creative capacities and communication skills remains to be seen.
The additional costs of smaller classes in a higher education system that is already viewed to be too expensive are clearly recognized. Nonetheless the self-reported learning benefits and positive attitudes toward smaller classes should not be ignored. Although our data are based on student self-report, many of the findings noted above merit testing using direct measures of student outcomes. At the very least, having a better understanding of the qualities of small and medium classes that support greater learning might improve the effectiveness of larger classes.
Steve Benton is senior research officer at the IDEA Center and emeritus professor of educational psychology at Kansas State University. Bill Pallett is the former president of the IDEA Center.