Support for policies that address climate change depends on an educated populace and its comprehension of difficult scientific concepts. To forestry action on climate change, the government of the United States in 2017 removed hundreds of webpages about climate change from the websites of federal agencies and departments and scrubbed the term “climate change” from thousands of others. Only four years later after a new administration took office was this censoring reversed. Also troublesome is that during this period some reliable sources of information became less suitable for educational purposes; for example, the Assessment Reports of the United Nations Intergovernmental Panel on Climate Change (IPCC) grew exponentially: the reports for Working Group I about the Physical Science expanded from 414 pages in 1990 to 3949 pages in 2021, for Working Group II about Impact and Adaptation from 296 pages in 1990 to 3675 pages in 2021, and for Working Group III about Mitigation from 438 pages in 1995 to 2913 pages in 2022 (Fig. S1). To address these issues, the National Science Foundation of the United States, as part of DUE 09-50396 “Creating a Learning Community for Solutions to Climate Change”, funded establishment of a nationwide cyber-enabled learning community to develop web-based curricular resources for teaching undergraduates about climate changes. One product of this project was a multi-disciplinary, introductory online course that is freely available to the public1.
This course was pressed into broader service as schools struggled to provide online materials at the onset of the COVID-19 pandemic. Institutions of higher education received criticism for adopting such courses, largely based on the assumption that online instruction is inherently inferior to that delivered face-to-face. The issue has become whether the convenience and safety of online instruction outweighs the possibility of inferior learning outcomes for today’s undergraduates.
Although the pandemic infused topical urgency into this issue, it is hardly new. The efficacy of distance learning has been debated since the External Program of the University of London first offered a correspondence course in 1858. Correspondence degrees have historically been driven by equity concerns for working people and women who could not access collegestwoyet they have historically been perceived as inferior to on-campus education3.4.
Online learning opportunities experienced explosive growth with the advent of widespread internet access and expanded credentialed university programs. In the United States alone, enrollments in online college courses rose from 1.6 million students in 2002 to 6.9 million students in 20185.6. During 2018, 35.3% of undergraduates in the United States took at least one course online, and half of these students took online courses exclusively6. This boom in online offerings coevolved with active learning and EdTech, and today’s online courses tend to be highly interactive, even when asynchronous or self-paced. Indeed, instructional design proponents often position today’s online courses on a spectrum with hybrid learning and flipped classrooms, rather than emphasize their ascent from didactic-style correspondence courses. Advocates for interactive online learning claim that a well-designed online course can be as effective as a face-to-face course, and perhaps even more effective than a traditional course based on passive lecture presentations.7,8,9,10.
Despite the new pedagogic paradigm for today’s online courses, familiar critiques of online learning persist4. Detractors cite high attrition rates as evidence that online courses leave students vulnerable to distraction and claim that the quality of educational experience and achievement in an online course cannot match that of a similar face-to-face class. Compounding these critiques, a number of studies in higher education have suggested that online courses, like their historic distance-education counterparts, tend to disproportionately enroll underserved students: if these already-vulnerable students are being attracted to a lower-quality educational experience online, then the proliferation of these courses might constitute an educational trap, exacerbating achievement gaps and providing barriers to persistence and success11,12.
The efficacy of online versus face-to-face courses seems ripe for an evidence-based study, yet high quality pseudo-experiments that compare the efficacies remain elusive. For example, The US Department of Education in 2010 conducted a meta-analysis of 28 studies comparing online versus face-to-face learning in post-secondary education settings and concluded, “When used by itself, online learning appears to be as effective as conventional classroom instruction, but not more so”13. A re-evaluation of this meta-analysis, however, found only four of these studies used an appropriate experimental design and examined semester-length college courses: in three of the studies, the students in the online versions of a course had poorer outcomes than those in the face-to-face versions, whereas in the fourth study, the students in the two versions had roughly similar outcomes14.
More recently, several large-scale studies of college students in the United States determined that student outcomes—both persistence through the end of the course and final grades—were substantially poorer for online courses than for face-to-face courses.15,16,17,18,19,20,21,22,23,24. These studies, however, were based on comparisons of courses with either different subject matter, those taught by different instructors, or those having relatively small numbers of students. Because many of these studies are based on dissimilar courses, they have had no opportunity to isolate students’ enrollment decisions to a simple choice between an online and a face-to-face version, nor provide appropriate analysis to account for the potential effects of underserved groups’ preference for one format over the other.
It follows that prior pseudo-experimental studies have also been unable to examine the critical concern that underlies all comparisons of online and face-to-face courses: if a tradeoff does exist between a face-to-face course’s baseline educational outcomes and an online course’s extended accessibility, is the decrease in learning outcomes worth the attendant increase in accessibility? These tradeoffs have been imbued with new urgency because of the COVID-19 pandemic during which universities and students seek to make difficult decisions about how to ensure safe course access while optimizing learning outcomes during the disruption of unfettered public life.
In this study, we seek to dissect student choice, student outcomes, and the tradeoffs between online and face-to-face courses at a large research university, through a post-hoc pseudo-experiment. We analyzed student performance versus their attributes for 1790 undergraduates of the University of California at Davis (a public research university) who enrolled in either an online or face-to-face version of the introductory course about climate change (for a syllabus of the course). see Table 1). Each demographic group had more than 100 students enrolled in the online and face-to-face versions (Fig. 1). Each year, both versions of the course were taught by a single instructor, thus, minimizing major confounding variables such as instructor bias, course design, content differences, and other aspects that might influence student choices and outcomes. Before the COVID-19 pandemic, we offered both versions of the course during eight Winter quarters and offered only the online version during six Spring quarters. In Winter and Spring quarters 2021, during the pandemic, we offered the course only online. For two concurrent course offerings in Winter 2019—one face-to-face and one online—and for COVID-19 pandemic-induced online course offerings in Winter 2021 and Spring 2021, we surveyed the students about their past experiences with online learning and how these experiences influenced their choice between the online and face-to-face versions of the course.
All elements of the course are available for free at https://www.climatechangecourse.org/, including a free multi-media textbook at https://indd.adobe.com/view/7eafc24d-9151-4493-85d2-cb3f2e5a2a51 that is updated regularly. During the period from 2017 to 2021, the online textbook had 5000 new users per year, who each averaged at least 3 views and 10 min per view. Before 2017, a printed version of the textbook was available for purchase25.