A Framework for Evaluating Class Size in Online Education Free

Higher education institutions are facing revenue challenges resulting from domestic and global economic forces. Even before the onset of the widespread recession in the United States, public universities saw cuts in state and county funding for education. Private colleges were not immune; many families with collegeage children are adjusting to lower incomes. As a result, academic administrators in virtually all university settings are seeking new sources of revenue (Alexander, 2006; Drago & Peltier, 2004; Jaschik, 2010; Kelderman, 2011; Keller, 2009; Labi, 2009; Potter, 2003; Schrecker, 2010; Walters, 2006).

Given the absence of need for on-campus classrooms and facilities in online courses, many academic administrators have looked to increasing the enrollment size of distance programs as a means of generating additional revenue. Meanwhile, faculty have generally found that the workload associated with some online courses expands with numbers of students. When enrollment numbers swell in courses, students experience less direct individual contact with faculty, and many faculty perceive that the quality of education declines as they have less interaction with students and a limited ability to engage with individual learning needs. Many studies report on faculty concerns about the quality of education offered online, both within their own educational institutions and also, more broadly, in the national and international arenas (e.g., Dykman, & Davis, 2008; “Forum,” 2010; Keeton, 2004; Legg, Adelman, & Levitt, 2009; Little, 2009; Parry, 2009; Schifter, 2000; Shea, 2007; Wickersham & McElhany, 2010).

With rising pressures to increase university revenues and a concomitant growing alarm among faculty asked to enroll more students and teach larger sections, the authors sought research guidance on the evidence surrounding online class sizes and educational quality. We found that definitive answers do not exist, but that there are relevant variables to guide determinations of appropriate class sizes (e.g., Drago & Peltier, 2004).

Research on online education is a multidisciplinary endeavor. Published studies of online research can be found in many journals, shown in Table 1, as well as in books written by online experts (e.g. Armstrong & Fukami, 2009; Hiltz & Goldman, 2005). We drew on all of these resources in our literature search, reviewing articles written for many different disciplines.

We sought to identify, based on accumulated evidence, the impact of various enrollment numbers on student learning outcomes, faculty workload, student satisfaction, and other related issues of concern, and, to identify factors that may help determine enrollment numbers in course sections without compromising learning. As part of our extensive, albeit nonexhaustive, inquiry, we searched for measurement tools and/or evaluative criteria to assess student learning resulting from varying class sizes.

Research conclusions on appropriate class sizes—number of students per faculty member—are mixed and, in some cases, contradictory (see summary in Table 2). Additionally, there is scant distance education research on student learning outcomes associated with varying teaching approaches. The effects of course size studied by a range of disciplines have yielded the predominant finding that varying class sizes will create different group dynamics and effects on faculty-student relationships (e.g., Arbaugh & Benbunan-Fich, 2005; Dykman, & Davis, 2008; Schellens & Valcke, 2006). Because of the immaturity of the research, the complexity of online education, and the number of relevant variables to consider (e.g., type of course, level of course, nature of learning expected in the course, setting, presence or absence of technology support and teaching assistants, faculty expertise in online education, evidence of student learning, etc.) (M. G. Moore, 2000), at this time the research findings on student:teacher ratios are inconclusive.

As shown in Table 2, we found studies that recommended anywhere from four to several hundred students in online courses. Larger classes are recognized as more impersonal and less individualized. Several studies have found that larger class section sizes were negatively associated with student learning, although “larger” was defined differently across studies (e.g., Arbaugh & Benbunan-Fich, 2005; Buckingham, 2003; Burruss et al., 2009; Schellens & Valcke, 2006). Research results indicate that classes larger than 30 students per faculty member can be delivered online, but such classes will inevitably take on the characteristics of one-way faculty-to-student communication unless the group is broken into smaller discussion sections with a faculty member or teaching assistant assigned to each small group and thereby creating, in effect, smaller class sizes. For courses that have traditionally been designed for small enrollments, such as seminars, because learning depends on ongoing faculty-student interaction, 20-30 students or fewer per group is recommended by some researchers, while others recommend sections sizes in the range of 15-25 students. There is little support for extremely small class sizes, such as, 3-10 students (Dykman, & Davis, 2008), except in doctoral education.

The online literature consistently indicates that online education benefits student access but is not more efficient; that is, the workload and intensity of effort for faculty are in general heavier for online education than for classroom-based education (Ascough, 2002; Drago & Peltier, 2004; Fjermestad, Hiltz, & Zhang, 2005; Parry, 2009). There is consensus that the single greatest predictor of positive selfreported student learning is instructor-student interaction. Teacher immediacy (timely and personal responsiveness) is one of the key drivers of student satisfaction (Bonnel, Ludwig, & Smith, 2008; Keeton, 2004; Schutt, Allen, & Laumakis, 2009). Student-to-student interactions/activities are also predictive of reported learning, but at a level half that of instructor-student interaction (Bernar et al., 2004; Keeton, 2004; Marks, Sibley, & Arbaugh, 2005). A third driver of student learning and satisfaction is ease of use of technology.

Because study findings are limited and results are mixed (e.g., Drago & Peltier, 2004), specific research-based determinants of class size are not sufficient at this time to support a standardized enrollment recommendation for all online courses (Hewitt & Brett, 2007). However, studies to date are suggestive, and from these it is possible to develop guidelines on class size deriving from the existing research.

A substantial body of distance learning research has developed in the past 10 years, informing online educators of effective teaching methods in web-based courses. Journals referenced earlier in Table 1 contain a plethora of studies identifying best practices. The goals and structure of quality online education are also well-researched and standardized through organizations such as Quality Matters (n.d.) and the Sloan Consortium (n.d.) (Little, 2009; Moore, 2008).

Determinations of pedagogically-sound student:teacher ratios relate directly to three dominant educational frameworks that appear repeatedly in the research literature: the objectivist-constructivist continuum, community of inquiry model, and Bloom's taxonomy. We discuss these below, followed by discussion of the implications of each for determinations of online class sizes.

The objectivist-constructivist continuum is a well-established construct in education (Legg, Adelman, & Levitt, 2009), differentiating two approaches to education. In the objectivist model, students learn passively by receiving and assimilating knowledge communicated to them by the professor. It uses largely one-way communication, and students learn individually, independent from others. Courses of a factual or scientific nature are effectively delivered by the objectivist method. Constructivist approaches assume that learning of new content results from complex interactions among individual students, faculty, and student peers. As students experience new information, they “compare this to internalized knowledge constructs based on past experiences, and then modify their constructs accordingly … knowledge has to be discovered, constructed, practiced, and validated by each learner” (Benbunan-Fich et al., 2005, p. 21). Constructivist environments offer multiple representations of reality and encourage thoughtful reflection on students' own and others' experience (Arbaugh & Benbunan-Fich, 2006, p. 436). Constructivist learning often requires that students break down, restructure, and transform existing knowledge to mediate understanding (Legg et al., 2009). Constructivist educators approach teaching with the belief that knowledge must be actively created or constructed by students and integrated with preexisting knowledge (Bain, 2004; Benbunan-Fich et al., 2005; Schellens, & Valcke, 2006). All university faculty select teaching approaches that fall somewhere on the continuum between transmitting knowledge to students unidirectionally, and engaging with students to reconstruct knowledge and make new information individually meaningful.

In online education, the choice of teaching method along the objectivist-constructivist spectrum should have a direct relationship to the number of students enrolled in a course. Faculty using predominantly objectivist approaches, and objective test-based or quantifiable assessment methods, can effectively teach very large numbers online. This form of education is teacher-centered. Students will generally learn equally well if they are in a class of five or 500. While the workload for faculty will expand somewhat with rising numbers of students, it does not increase directly with class size. Conversely, constructivist teaching approaches require much smaller class sizes because they are learner-centered. The student work of learning—deconstructing old knowledge and integrating new and more complex information—typically depends on faculty interaction with individual students, and regular assessments and feedback; it is student-centered and teaching-intensive. Indeed, schools and universities exist because few people are capable of robust learning without focused facilitation from knowledgeable experts. Faculty workload using constructivist teaching methods expands directly in relation to the number of enrolled students.

In objectivist learning, our literature review revealed that there is no recognized upper limit to the number of students per faculty member enrolled in online classes. Since learning occurs with the transmission of content from faculty to students in a teacher-centered process, and each student is expected to learn on his or her own, reflected in their performance on objective fact-based assessments, courses may enroll as many students as is logistically feasible. On the opposite end of the objectivist-constructivist continuum, smaller course sizes are necessary when there is a need for faculty interaction with students to assure learning. Faculty delivery of content, direct instruction, interaction with students, correction of misconceptions, and formative and summative evaluations of student learning constitute time-consuming pedagogy. As reflected in Table 2, a preponderance of the evidence suggests that no more than 20 students should be enrolled in such constructivist-designed courses, and some researchers argue for even smaller numbers.

In teaching-intensive constructivist learning classes, the research indicates that the numbers of students are a significant driver of faculty workload. In courses that use a combination of objectivist and constructivist approaches to teaching, considerations of student learning and faculty workload make it a judgment call for what the “right” number of student enrollments should be. Decisions on courses that fall in the middle of the objectivist-constructivist continuum need to take into account both learning effectiveness (largely a faculty determination), faculty workload, and university revenue needs (largely an administrative determination).

First developed in 2000 by Garrison, Anderson, and Archer, and later supported by the results of numerous studies (e.g., Anderson, Rourke, Garrison, & Archer, 2001; Arbaugh, 2007; Arbaugh & Hwang, 2006; Garrison, 2007; Garrison, Anderson, & Archer, 2010; Garrison, Cleveland-Innes, & Fung, 2010; Meyer, 2006; Shea, 2006; Shea, Li, Swan, & Pickett, 2005), the community of inquiry (COI) model in online education assumes that the instructor's role is critical in potentiating student learning. The model proposes that teaching, cognitive, and social presence all contribute significantly to learning effectiveness within an online community.

Teaching/teacher presence in online education involves the design, facilitation, and direction of learning in service of students' construction of meaningful and educationally worthwhile knowledge. Teaching presence is conceptualized as instructional design and organization, facilitating discourse and building understanding, and direct instruction (see Table 3). Activities associated with teaching presence are extensive, ranging from course design and syllabus construction, to course and learning strategy methods, to regular faculty interactions with individuals and groups of students, to formative and summative feedback. Cognitive presence marks the extent to which students demonstrate construction and integration of new meaning through sustained learning activities. Knowledge construction by individual students is often more visible in online courses than in classroom-based courses. Cognitive presence is influenced by the faculty's teaching and social presence, and by other students' cognitive and social presence. Social presence is reflected in the ability of faculty and learners to project themselves socially and emotionally into a course, and create an identity as a “real person” in the online environment. It too is affected by the faculty's teaching and social presence, and by other students' cognitive and social presence.

Beginning with the teacher's design and management of a course and his or her teaching presence, comprehensive use of the COI model has been shown to enhance student learning and satisfaction (Anderson et al., 2001; Arbaugh, 2007; Brook & Oliver, 2003; Garrison, 2007; Garrison et al., 2000; Meyer, 2006; Paulus et al., 2010; Richardson & Swan, 2003; Swan & Shea, 2005; Swan & Shih, 2005). Within the COI framework, teacher immediacy is a recognized driver of student learning and satisfaction. Immediacy “refers to behaviors that lessen the psychological distance between communicators” (Swan & Shea, 2005, p. 242), and can include a range of faculty actions, such as prompt and focused replies, warmth and friendliness, addressing students individually, use of humor or emotion, self-disclosure, greetings and closures, and use of connecting language (Lahaie, 2007).

Promotion of critical thinking is an additional element of the COI model especially relevant to higher education. Many educators find that active online learning strategies enhance faculty access to students' thinking capabilities (Leppa, 2004). Online teaching presence provides students with intellectual and scholarly direction by sharing faculties' subject matter expertise and assisting students with knowledge constructions, especially application and integration. By diagnosing and correcting misconceptions and providing students with feedback, faculty presence diminishes students' internalized barriers to learning new information.

Consistent with the COI model, when faculty prompt students to project social and cognitive identities into online education, learning assumes a more constructivist than objectivist profile. Objectivist teaching does draw on many elements of teaching presence, especially instructional design, organization, and direct instruction, but diminishes the relevance of social presence in one-way instruction. Cognitive knowledge is largely experienced autonomously by, rather than shared among, individual students.

Bloom's taxonomy is the third proposed educational framework related to determinations of appropriate class size (Bloom, Englehart, Furst, Hill, & Krathwohl, 1956). The taxonomy is a classification of 6 levels of learning, moving from lower levels to higher-order thinking: knowledge, comprehension, application, analysis, synthesis, and evaluation. Faculty design instruction to draw on various levels of Bloom's taxonomy, and subsequently assess students' learning outcomes according to educational intent (Meyer, 2004). Although there is considerable variability in targeted taxonomic levels in higher education, typically more basic knowledge and comprehension levels are addressed in lower division courses in college, with more complex learning and critical thinking expected in upper division and graduate courses. Doctoral study disproportionately draws on learning at the analysis, synthesis, and evaluation levels. Arend (2007) also describes student learning associated with basic to complex cognition hierarchically, but uses different terminology: rehearsal (memorizing, reciting), elaboration (building internal connections), organizational (selecting appropriate information to connect to new information), critical thinking (developing new ways of thinking about content, applying prior knowledge to new situations, reaching decisions, and making evaluations), and metacognitive self-regulation (identifying how to control and modify individual cognitive processes).

By historical practice, universities implicitly recognize that greater faculty-student interaction is necessary for effective education at the upper levels of Bloom's taxonomy: undergraduate courses tend to be large, sometimes with hundreds of students enrolled in introductory courses; upper division courses for students in their major move toward medium sizes of 20-50 students per faculty member, with some seminars of 20 or fewer students. In graduate education, the majority of classes are 25-30 students or fewer, often providing small seminars the final year; and doctoral level courses are almost always small seminars consisting of 12 or fewer students. These course sizes reflect common knowledge—that higher-order thinking and learning require more intensity of student-faculty interaction. In parallel form then, online education courses should reflect larger classes at lower levels and smaller classes at higher levels of Bloom's taxonomy.

Were all college courses readily categorized as to their needs for objectivist or constructivist teaching strategies, targeted as to required levels of learning in Bloom's taxonomy, and consistently implemented using the COI model, it would be relatively simple to develop standardized enrollment sizes for online courses (see Table 4). But none of these constructs is consistently quantified and applied to college courses by discipline, level in the curriculum, type of student, or faculty preference. Courses with large amounts of factual information, such as introductory biology, may tend toward objectivist design, a need for teaching presence but little social presence, and student learning at the knowledge and comprehension levels. More advanced courses, in contrast, such as analysis of literary works by historical period, may require constructivist learning strategies, a high level of teaching and cognitive presence, and learning at the application, analysis, and/or synthesis levels. Teaching intensity would be far greater for the latter course than the former, and thus smaller student section sizes would be appropriate to assure learning. Graduate courses in statistics or physiology may have a mix of teaching intensity factors and be appropriatelysized at medium enrollments of 20-30 students, whereas a research methods and design course could demand greater faculty interaction with students and tend toward the more teaching-intense elements of these frameworks. Doctoral-level courses and dissertation advisement are invariably high on constructivism; teaching, cognitive, and social presence; and at the upper levels of Bloom's taxonomy, thereby requiring very small numbers of students per faculty member (see Table 4).

The majority of college courses at the undergraduate level, and some courses at the graduate level, fall in the vast “in-between” gray areas of the three frameworks, making it impossible to standardize online course size. A political science curriculum may initially call for learning at the lower three levels of Bloom's taxonomy, but eventually demand constructivist teaching strategies to assure student assimilation of new ideas into preexisting knowledge constructs. A masters- or doctoral-level course on research methods may effectively use objectivist teaching strategies for basic concepts (e.g., research design and methods) but require interspersed constructivist COI modules to assure mastery. A doctoral-level course may be taught at higher levels of Bloom's taxonomy while using objectivistdelivered criteria for evaluating research articles. Thus, when the three frameworks are used with a mix of high and low teaching intensity factors, there is a need for a collaborative judgment call, with faculty serving as pedagogical experts and academic administrators weighing revenue pressures. For a mix of teaching-intensive and teaching-light elements in courses, it is useful for institutional parties to identify a priori the parameters of course enrollments sizes (e.g., 18-28 students). High quality pedagogy and financial stability are each central aspects of healthy universities and both must be addressed in decisions about course enrollment.

Online education has come of age and its future is bright. Distance learning is well established in higher education domestically and internationally, and it has enabled universities to educate more diverse and dispersed populations of students than was possible in location-specific institutions. With maturation, online education faces a rising call to examine and improve the quality of teaching and learning from a distance.

We have presented here the state of the evidence surrounding one factor related to quality of education: enrollment sizes in online courses. The current research findings on the “right” size of online courses is instructive but not conclusive. Three dominant educational frameworks—the objectivist-constructivist continuum, COI model, and Bloom's taxonomy—were presented to provide guidance in determining pedagogically effective and revenue-sensitive online course sizes. The frameworks provide insight about teaching-intensive or teaching-light online courses, and serve as key factors for understanding the educational and workload dynamics of online teaching. Based on the existing research, we have suggested ranges of course sizes associated with higher or lower levels of teaching intensity (see Table 4).

There are a number of pedagogical and technological paths forward that will advance our understanding and improve the quality of online education. For those aspects of teaching and learning associated with the faculty role, one path points toward the development of research-based measures and evaluative criteria to assess student learning outcomes within courses of varying class sizes. The three educational frameworks presented above are likely to play prominent roles in these developments.