Higher education is faced with many challenges at this time. Two, however, stand out as providing critical tests for the future viability of many institutions. The first is the growing set of constraints on revenues, and the second is the increasing necessity to improve student learning significantly (How can we think about the wave of new innovations in higher education?). Figlio, Schapiro, and Soter (FSS) recently published an important National Bureau of Economic Research working paper whose provocative title I have borrowed for the title of this post. That paper has important implications for responding to both of these challenges.
Analyzing data for eight cohorts of first year students at Northwestern University, FSS conclude:
We find consistent evidence that students learn relatively more from non-tenure line professors in their introductory courses. These differences are present across a wide variety of subject areas, and are particularly pronounced for Northwestern’s average students and less-qualified students.
In a highly oversimplified explanation, this result was obtained by showing that students who took the first course in a sequence of two courses from a non-tenure line faculty on average did better in the second course in the sequence than did those students who took the first course from a tenure line faculty. The actual experimental protocol and data analysis was, of course, much more complex than this oversimplified explanation might suggest.
This result has been widely quoted in various media, usually in rather dramatic form (e.g. Study gives tenured professors a failing grade). While I agree that this was indeed a dramatic result, I believe this study is only a excellent early step in an area of research that could have more impact on the future of higher education than the much-discussed MOOCs.
In the next section, I will look a bit more deeply than has been the norm in the press thus far at what this study does and doesn't do, and where it fits within a general area of research on improving student learning. In the final section, I will describe why I believe this area is of such key importance for higher education.
This is certainly not the first time that research that focused on measuring student learning outcomes has demonstrated that tenure status of the teacher is not a significant variable. For example, one of the most striking recent results was described in Science by Deslauriers, Schelew, and Wieman (DSW). In this experiment, Deslauriers, a postdoc in the University of British Columbia's Carl Wieman Science Education Initiative, and Schelew, a masters student in physics at UBC, took over the teaching of a subject unit in one of the two sections of a large introductory physics class. They taught their students using interactive activities based on research in cognitive psychology. The other section was taught using a traditional lecture by a highly rated professor with many years of experience. The test at the end of the unit showed almost no overlap in distributions of scores associated with the two sections: the students taught using advanced pedagogical approaches had learned almost twice as much as the students taught in traditional ways. Related research done at the University of Washington in a biology class described similar outcomes, and pointed that educationally disadvantaged students benefited particularly from the new pedagogical approaches. In these and essentially all other related studies, the significant variable has been the pedagogy used, although in practice the non-traditional teaching very often has been carried out by non-tenure line personnel who spent considerable time gaining expertise in advanced techniques.
The FSS study is obviously done in a very different way than most of the previous research in student learning. FSS has the benefit of taking a broad, high-level view of the problem. It did not focus on one single course or part of a course, but on the first year educational outcomes for all entering freshmen at Northwestern for an eight year period. It identified that teacher status was important in student learning in a very wide variety of types of introductory courses, but it did not try to look at the mechanisms - the "why" - behind this dependence on teacher status.
The "why" is, of course, critically important to understand. Functionally, previous data indicate that magnitude of learning is related to the pedagogical approach used, and that the newer approaches generally are much more effective than the older. Were the non-tenure line faculty on average using a more sophisticated teaching approach than the tenure-line faculty? This is possible, since non-tenure track teaching faculty are rewarded for teaching, while tenure line faculty are rewarded for both research and teaching (usually with emphasis on the former). Or is there some additional variable at play, e.g. instructor viewpoints more closely related to practical implications rather than more theoretical academic concerns.
This study is similar to most (but certainly not all) other studies of student learning in that it focuses on introductory classes. Thus, even if we accept that "tenured professors get a failing grade" for the teaching of introductory courses, we lack information about more advanced courses. At one extreme, it seems probable that for very advanced courses that deal with latest developments in a field (e.g. many graduate courses), the potentially more specialized and up-to-date knowledge possessed by a research-active tenure line faculty would reverse the findings. In between the introductory courses and the very advanced courses is a very large gulf. Thus a critical focus for additional research would be to define the characteristics of courses in which use of non-tenure line teachers is likely to lead to better student learning than would be obtained using tenure line teachers.
Another important direction for additional research would be to repeat this experiment in other institutions of different types. Northwestern is one of the nation's premier research institutions, and its non-tenure and tenure line faculty differ in many ways from similarly designated faculty at many institutions having a different focus and mission. In particular,
Almost all classes taught by non-tenure track faculty at Northwestern are taught by
those with a longer-term relationship with the university.
The fraction of faculty nationwide who are in non-tenure lines is rapidly growing across all institutional types, and in many institutions a large fraction of these faculty are part-time and/or do not have a continuing relationship with the institution.
FSS note that some earlier studies comparing use of non-tenure vs tenure line faculty in teaching have shown the importance of tenure-line faculty in such critical outcomes as graduation. However, many of these studies do not differentiate between adjuncts who have continuing relationships and/or are full time, and those who do not. My prediction is that this is a critical variable in understanding the role of tenure in student learning.
In How can we think about the wave of new innovations in higher education? I discussed the business model for higher education introduced by Christensen, Horn, Soares, and Caldera (CHSC). Although there are an infinity of possible different business models, CHSC pointed out that there are only three generic types of business models. Further, business research across a wide spectrum of industries had shown that an enterprise that simultaneously utilized more than one of these generic types incurred very high overhead costs. The is bad news for the very large slice of higher education that has a mission that involves teaching, research and student social growth: each of these three mission elements is carried out using a different generic type of business model. Indeed, this simultaneous use of all three generic types of business model was identified by CHSC as the underlying basis for the very high cost of much of American higher education.
Another, less emphasized drawback of using multiple business models within one activity is that no one model can be optimized: each must operate under constraints set by the others. As a consequence, each component will be suboptimized both in the quality and quantity of its output and in its cost of production. In a research university, the research function is favored in the optimization process, while in a college, the teaching function is favored. In neither case, however, can real optimization of either component occur.
In the traditional American research university, much of this overhead and suboptimization comes from having the two distinct functions of teaching and research carried out by an individual - the faculty member. Visible research active faculty move in a global marketplace, which makes their salaries much higher than teaching faculty who tend to move in a more localized marketplace. Because they spend much of their time doing research and consequently carry a lower teaching load than the typical teaching faculty, this high salary converts to a much higher cost per student than is the case for teaching faculty. Since promotions and reputation are built primarily on research output, it is unusual for a research faculty to spend the time to learn new, more effective pedagogical approaches and to develop expertise in them. Thus the teaching function is suboptimized both in effectiveness and cost when a single individual does both research and teaching. On the other hand, the time spent teaching plus all of the teaching-related committee activities greatly decrease the time available for research, and often make it difficult to create the periods of unstructured time that enable research to flow most smoothly. Thus, the research function is not optimized as well.
The importance of the research of FSS, DSW and others is that it demonstrates that in some circumstances the interlocking multiple business models of research and teaching can be separated, leading simultaneously to improved outcomes in both the teaching and research functions. This information is very likely provide a key component of the long-term survival strategy of many of the over two hundred research universities in the US.
One can argue that higher education is already moving in the direction of separating teaching from research, since the percentage of non-tenure track faculty is increasing in almost every sector. However, this change so far seems generally to be driven by financial constraints - half of the problem - and little attention is being paid to quality of learning - the other half of the problem. By doing this move strategically using data driven learning research, it should be possible to make real progress on both halves.