Online Education and the “Cost Disease”

By Henry Reichman

Higher Education in the Digital Age by William G. Bowen in collaboration with Kelly A. Lack. Princeton, NJ: Princeton University Press, 2013.

Higher Education in the Digital Age
by William G. Bowen and Kelly A. Lack

Former Princeton University president William Bowen, who has done important, even courageous, work on race and affirmative action and on the insidious problems in college sports, returns in this small (161-page) book to a theme that has long absorbed him: the seemingly inexorable increase in the cost of higher education. Higher Education in the Digital Age consists of edited and expanded versions (with extensive endnotes) of two lectures he delivered at Stanford University last year on the potential of online teaching to help control the rising cost of higher education. These lectures are accompanied by comments from Harvard professor Howard Gardner, Stanford president John Hennessy, Columbia professor Andrew Delbanco, and Stanford professor Daphne Koller, who is also cofounder of the for-profit online provider Coursera.

While the book offers a brief and readable introduction to some issues in online education (a five-page appendix, “The Online Learning Landscape,” is especially helpful, as are the endnotes), it is most useful as a largely unintended demonstration of the unconscious myopia of leaders and some faculty members at elite institutions like Stanford. With the notable exception of Delbanco’s eloquent and pointed remarks, the authors seem largely ignorant of how faculty teach and how students learn at the overwhelming majority of colleges and universities in the United States, which Bowen repeatedly and dismissively refers to as “less selective institutions.” The authors draw examples largely from their own universities or from other research institutions. They make little effort to address the problems faced by working and commuting students, by those with disabilities, by students whose high schools were substandard, and by those who stand on the have-not side of the “digital divide.” The entire book includes just three passing references to community colleges, by far the largest and fastest-growing segment of American higher education, where issues of cost control and online learning are critical.

Bowen’s first lecture addresses what he has previously named the “cost disease.” The term refers to the proposition that “in laborintensive industries such as . . . education, there is less opportunity than in other sectors to increase productivity. . . . As a result, unit labor costs must be expected to rise faster in . . . education than in the economy overall.” In other words, according to Bowen, the principal driver of cost inflation in higher education is the inevitably rising price of faculty and the inability to replace faculty with less costly inputs.

This review is not the place to examine this thesis in depth, but there is much to suggest that whatever its theoretical appeal it seems not to be borne out by reality. Indeed, the Delta Project on Postsecondary Costs, Productivity, and Accountability has labeled the idea that faculty costs drive cost inflation a “myth.” As researchers Dennis Jones and Jane Wellman demonstrated in their article “Rethinking Conventional Wisdom about Higher Ed Finance,” “spending on faculty is a minority of total spending in most institutions, a proportion that has been declining in all sectors for the last two decades.” According to Delta’s figures, spending on instruction, mainly faculty salaries, in public master’s-granting institutions, for example, amounted to just 34.3 percent of total spending in 2010. In private research institutions the figure was only 30.0 percent.

Take one example: The California State University system, the country’s largest four-year institution, enrolls more than 400,000 students. For the 2011–12 academic year, the CSU reported $7.4 billion in expenses, of which only $2.3 billion, or 30.5 percent, went to instruction, including faculty compensation. In the CSU, full-time equivalent enrollment increased by 33.6 percent from 1990 to 2012. But FTE faculty numbers actually declined by 2.1 percent during that period, 11.5 percent for tenure-track faculty. Yet from 1993 to 2010 the number of administrators and senior staff jumped by a whopping 51 percent.

While Bowen recognizes that expenditures on faculty are but one factor driving costs up, Hennessy puts the argument more baldly: “Faculty salaries have gone up faster than wages for most other workers,” he declares, adding that for those in the academy, “financially, life is much more pleasant now than it was thirty years ago.” Maybe at the Princetons and Stanfords, but not elsewhere. According to the AAUP’s Annual Report on the Economic Status of the Profession, salaries of full-time tenure-track faculty at all institutions, adjusted for inflation, actually declined by nearly 3 percent from 1971 to 2012. More important, today less than one-fourth of all teaching faculty are on the tenure track. Among the remainder, the majority are either part time and “temporary” or graduate student employees.

It is notoriously difficult to obtain accurate salary information for such faculty, but based on limited data collected by the Coalition on the Academic Workforce and reported by the AAUP in its 2012–13 faculty compensation survey, even at private doctoral research institutions their median salary per course was just $3,800 in fall 2010. At public community colleges the median was only $2,250. From these figures it is difficult to sustain the notion that for such instructors, the majority of the higher education instructional workforce, “life is much more pleasant” than it was for their counterparts thirty years ago.

Remarkably, both Bowen and Hennessy suggest that another driver of increased cost has been the proliferation of PhD-granting institutions and consequent overproduction of doctoral degrees, which are more expensive to generate than bachelor’s or master’s degrees. Yet neither acknowledges how the glut of doctoral candidates and degrees permits research institutions to cut costs by exploiting graduate students as instructors (and, for research in the sciences, as inexpensive postdocs), while “lesser” institutions may dip into a growing oversupply of underemployed PhDs for “temporary” instructors. Indeed, their outrage seems motivated less by genuine concern about cost containment or attention to basic fairness than by thinly disguised resentment that some “less selective institutions” have the gall to compete with their own elite universities.

In his second lecture, Bowen turns to “Prospects for an Online Fix.” In 2000 he posited that “sound online instruction is likely to cost more than traditional instruction,” but he now concludes that conditions have changed. To be sure, his argument is hedged by numerous cautions; Hennessy and especially Koller are the true online evangelists. But Bowen is convinced that “greater access to the Internet, improvements in Internet speed, reductions in storage costs, the proliferation of increasingly sophisticated mobile devices, and other advances have combined with changing mindsets to suggest that online learning, in many of its manifestations, can lead to at least comparable learning outcomes relative to face-to-face instruction at lower cost.”

Asking whether online formats are good or bad makes as little sense as asking whether, say, the lecture or the seminar format is good or bad. The answer depends on where, how, and for and by whom. Certainly online education has become a commonplace feature on the educational landscape at all levels and has proven valuable to many students and faculty members. At the same time, however, as Bowen acknowledges, we have little hard data to demonstrate whether—and, more importantly, in which subjects and for which students—online instruction yields outcomes comparable to traditional methods. A 2010 meta-analysis by the Department of Education found only seven rigorous studies of fully online, semester-length courses at four-year institutions. These studies showed little difference in learning outcomes, but the typicality of the courses was unclear, and the studies did not investigate completion rates, which are frequently much lower online.

Bowen places considerable emphasis on one study of a hybrid statistics course offered by Carnegie Mellon University, which “found no statistically significant differences in standard measures of learning outcomes” from those of students in traditional classes. But statistics may well be a subject especially suited to online formats. More important, the hybrid (or blended) format, in which online material is combined with face-to-face meetings, is probably the format least likely to cure Bowen’s “cost disease.” Recently, a hybrid engineering class at San Jose State University offered in cooperation with the online provider edX attracted considerable attention when it was reported that students in the class succeeded at remarkably higher rates than those in traditional sections. But these “online” students met face-to-face with instructors three hours each week, and their instructors acknowledged spending significantly more time preparing and evaluating student work than they did in traditional classes.

Perhaps the most extensive study of online outcomes was conducted by the Community College Research Center at Columbia University’s Teachers College. Researchers there assessed the experience in fully online courses of multiple cohorts totaling more than fifty thousand students at twenty-three community colleges in Virginia and over one hundred thousand students at thirty-four community and technical colleges in Washington State. The findings were sobering. For one thing, withdrawal rates were consistently nearly double in online courses what they were in face-toface classes. Moreover, regardless of academic subject, demographics, or academic background, investigators found that “the same student performs more poorly in a fully-online course than in a face-to-face course.” Students with a stronger academic background showed a small dip in performance, while more poorly prepared students evidenced a larger decline. Finally, performance gaps between white and minority students tended to widen in online courses.

In short, whatever online teaching’s virtues, it is by no means clear that comparable learning outcomes and completion rates can be sustained online for all students, especially students most “at risk,” or in all subject areas. And it is equally questionable whether online formats measurably reduce costs.

Of course, the online format with the greatest promise for cost reduction is the MOOC, or massive open online course. Bowen’s lectures were delivered as “MOOC madness” was first taking off, and his endorsement—especially when compared with the hype that has followed—is tentative. He argues, probably correctly, that MOOCs might save money only if they remain standardized and are not “customized” to a local student body or faculty preferences. This leads Bowen to argue for “genuinely collaborative decision-making that includes faculty, of course, but that does not give full authority to determine teaching methods to particular professors or even to particular departments.” In the past, he concludes, “it may have been sensible to leave almost all decisions concerning not just what to teach but how to teach in the hands of individual faculty members. It is by no means clear, however, that this model is the right one going forward.” Despite the obligatory nod to “collaboration,” such remarks come close to being a full-scale assault on fundamental principles of academic freedom and shared governance.

More important, however, is that what Bowen calls “customized” teaching formats are what most faculty members understand as the compelling need to address the reality that not all students learn in the same way. What may work for students at Stanford may not be suitable for those at San Jose State or a community college. If MOOCs amount to little more than canned video lectures combined with online exercises and discussion boards, they will surely fail. Indeed, the single-digit completion rates reported for most MOOCs should at least inspire greater caution about their long-term value.

Koller, however, argues passionately that online teaching “is really a new educational paradigm.” Teaching online, she claims, compels instructors to “flip” the classroom— “not the kind of thing that we were trained to do,” which was instead to “stand and orate.” The sheer parochialism of such a comment is astonishing. In most teaching institutions, faculty have for several decades moved well beyond the lecture format to modes of instruction that involve greater active student participation, be it face to face or online. “Be a guide on the side, not a sage on the stage,” we’ve been counseled repeatedly by administrators, faculty development officers, and many colleagues. Although interactive education may not be the norm at Stanford, it is quite common at institutions like the CSU, in community colleges, and elsewhere, perhaps even more so in traditional than in online classrooms.

For Koller it is virtually a matter of faith that online formats will facilitate a transition from passive to active learning while simultaneously reducing the cost of college and potentially offering a Stanford education to all. Bowen is more cautious, but his enthusiasm for these new formats also appears largely faith-based. One may hope they are correct, but it would be advisable to remain highly skeptical.


Henry Reichman is professor emeritus of history at California State University, East Bay. He is first vice president of the AAUP and chair of Committee A on Academic Freedom and Tenure. His e-mail address is [email protected].