In the struggle to find ways to reduce the costs of a college education, one of the greatest challenges has been understanding the drivers of costs and their value in terms of quality and outcomes.
Leaders in higher education—those of us in the middle of the cost debates and the ones most able to take direct action at our institutions—must push harder for better data and analysis on costs, quality, and, ultimately, student outcomes.
One way to begin teasing apart the nuances of the cost issue may come from the world of health care. In 1973, Science published a seminal paper about the problem of health-care costs, noting a simple finding: Health-care expenditures and the use of certain health-care services varied significantly across parts of Vermont, with no evidence of causes other than different practice styles. Because many of the areas’ populations were similar, it seemed clear that differences in health-care needs were not driving the differences in costs.
That study helped to accelerate the growth of the entire field of health-services and health-economics research in the United States, including the study of quality as a health-care measure and outcome, by illustrating the potentially huge expenditures wasted on services that did not have a clear benefit to the nation’s health. That still-growing body of research ultimately became the basis for a range of policy strategies to reduce waste and improve the quality of health care.
Simple comparisons among colleges may yield similarly valuable insights. Yet one challenge in comparing costs among this country’s 4,000 institutions of higher education is the wide variation in mission, size, research intensity, programs, and student population. The community of highly selective, private liberal-arts colleges may be one place to start in comparing costs, given that they have similar missions, similarly accomplished entering students, relatively little research infrastructure, and few, if any, graduate programs.
As president of one such institution, Grinnell College, I began to look at our costs against those of comparable institutions listed in the Integrated Postsecondary Education Data System, the federally mandated reporting system for American colleges. By costs, I refer to the cost to the institution—how much the college spends to provide education per student—and not the price charged to students who receive no aid. I restricted my comparison group to the private liberal-arts colleges with the 30 highest 75th-percentile combined SAT scores, excluding Grinnell. The six-year graduation rate for all but one college on the list was over 80 percent (the average was 90 percent, and the outlier was 74 percent). I used Ipeds data to estimate expenditures per enrolled, full-time-equivalent student.
What I found was eye-opening. In 2012, expenditures in this group varied from approximately $75,000 per student per year at the high end to $35,600 at the low end. That large variation raises the question: Is it really valid to assume that the quality of education at the highest-cost institution is more than twice as great as the education at the lowest-cost institution? Are the additional costs at the high end associated with any gain in terms of the educational quality or content—and, if so, is the gain sufficient to justify the added investment by students and their families? (The amount that Grinnell spends per student is approximately $50,500 per year, right in the middle.)
I would argue that the reason for the variation is unlikely to be simply a matter of regional differences in input costs (for example, that one area is more expensive to operate in than another). Even among clusters of colleges in the same geographic community there are substantial differences. In one such cluster of three institutions on the West Coast, the costs per student are $65,963, $58,405, and $51,725. In another cluster, of East Coast colleges, costs are $70,588, $59,417, and $51,967.
Based on my admittedly crude exploratory analysis, I hypothesize that the large variations in costs per student—even when in the same geographic region among colleges of a similar type—might be due to differences in educational “practice style,” similar to how the differences in health-care costs and use in the Vermont study were attributed to differences in health-care “practice style.”
What might be embedded in those different “educational practice styles”? What might be the marginal costs and gains in outcomes? Does the difference in cost between a college with a student-to-faculty ratio of 8:1, versus 9:1, versus 15:1 result in a meaningful (or, to borrow a term from health care, a “clinically significant”) difference in outcomes or quality? Is there a meaningful difference in outcomes for classes of 20 versus 25 students? Does the marginal impact of changes in class size vary by discipline? What role do size and composition of administrative structure play? What is their payoff in terms of outcomes? What is the impact of teaching and ancillary facilities on costs and outcomes?
Those are the kinds of questions that we, as leaders of colleges and universities, must investigate more fully. I believe it is time for us to develop a research agenda to help us understand our expenditures and their value in a more substantive way. If our institutions are to flourish and if we are to continue to receive broad public support for our missions, we must control costs without stripping our educational offerings of their essential value for achieving our missions—and right now we have limited data and analysis to inform how we do that.
We must identify the resources to support a “higher-education services” research infrastructure, from foundations and other private funds and from public sources, and we must be able to analyze the findings collectively. The obvious place to start is among colleges with similar business models, missions, and student populations. With time, we should be able to learn across different types of institutions as well.
Our primary goal should be to explore, in a meaningful way, how any potential changes we might carry out will affect outcomes, quality, and costs. Once we have those kinds of data and analysis, we need a method to disseminate findings quickly, in a format that is useful and easy to understand for those who make allocation decisions.
Finally, we need to motivate colleges to work together. If groups of educational institutions unite to function as “cost and quality cooperatives,” we can tackle those issues more quickly as a cohesive community rather than individually. Such cooperation and commitment may encourage the White House and Congress to propose more thoughtful, evidence-based incentive systems for controlling costs instead of the rudimentary policies proposed to date.
To trigger any of this, we must start challenging ourselves to ask the right questions, even if they are difficult. It’s time to follow the lead of the health-care community: Develop a strong body of evidence to guide incentive policies that will control costs and improve outcomes, with minimal unintended negative consequences.
