Chapter X

Grading Teachers:

Academic Standards and Student Evaluations

Lester H. Hunt

I sometimes entertain my non-academic friends by telling them that, at the end of each course I teach, before I compute my students’ grades, I pause nervously while I wait to be graded by my students. This process can be described less paradoxically, but surely no more truthfully, as follows. In my department, and as far as I know all the departments at my university, each course ends with students anonymously filling out forms in which they evaluate the teacher and the course. The form includes several questions in which the student is asked to rate the teacher in various respects (such as clarity and organization, availability outside the classroom, and so forth) along a numbered scale (in our case, from one to five); they are also asked one very general question in which they are told to rate the instructor's overall effectiveness. They are also invited to write comments on these matters. Mean scores (in effect, grades) are calculated for each question and published by the university. In addition, these student evaluations of teaching (often called SETs for short) are used each year by the committee that decides merit pay increases for faculty. When the faculty member is being considered for advancement to tenure or promotion to the rank of Full Professor, these evaluation forms are supplemented by faculty evaluation of teaching, in which faculty members visit the candidate’s classes and report on her or his effectiveness as a teacher. Except for these two once-in-a-lifetime events, the student evaluation forms are the only way in which we judge the quality of teaching. In other words, teaching is for the most part evaluated by students and not by faculty.

What I wish to do here is explain and defend the following thesis: that such evaluation forms, administered and used as I have just described, are a very bad idea. The only thing that could justify such a policy would be that all the alternatives are, all things considered, even worse. Among the most important considerations that support this thesis are the distorting effects such a policy has on pedagogical method, especially the effect we can reasonably predict it will have on academic standards and grading policies.

One clarifying remark is probably needed before I begin. There are two radically different purposes for which student evaluations can be used. One is their use, by departments and administrators, to aid in personnel decisions, such as pay raises and promotions. The other is their use by instructors who wish to improve their teaching. No one would deny that such evaluations have a role to play in pursuing the latter sort of purpose. I will not be discussing this function of evaluations here, except to point out that evaluation forms that are developed for personnel purposes are in general not intended to help the instructor to become a better teacher and would be poorly adapted to serve that end, compared to forms that are written for the express purpose of improving teaching effectiveness.1

Three Problems

As I see it, the current system, in which we rely almost exclusively on anonymous student forms in evaluating the quality of teaching for personnel purposes, raises three serious problems. I will call them a) the epistemological problem, b) the problem of academic standards, and c) the problem of the distortion of pedagogical styles.

A. The Epistemological Problem. First, I think it will shed some light on the relevant issues if we pause for a moment to try to imagine a system that clearly would, if it could exist, measure teaching effectiveness. We can easily conceive of such a system if we consider the way the earliest universities were organized. During the Middle Ages, professors offered lecture series on subjects that students needed to know in order to receive their degrees. The lecture series was just that: they were lectures. The professor did not evaluate the students in his own lecture course. Rather, the students were evaluated by taking examinations that were separate from the course of lectures.2 In such a system, it would be a fairly straightforward matter to measure the effectiveness of a professor’s teaching. One need only look to see how the students in a given Professor’s course performed on the corresponding examinations. In fact, an ideal method would be to examine the students on the relevant subject both before and after the lecture course, then correlate the degree of improvement (if any) with the instructor whose course the student attended.

The reason why this would be an ideal evaluation procedure is obvious. Teaching is a purposive activity, the purpose being (in some sense of these words) to impart knowledge. It is successful to the extent that it achieves this intended result. It is equally obvious why student evaluation of teaching effectiveness is epistemically problematic. One can only tell whether the result has been achieved if one already possesses it: The only judge of whether a person has come to know X-ology is someone who knows X-ology. Yet whether a given student is in the position of a competent evaluator – whether the student has come to grasp X-ology – is precisely what is at issue when we are evaluating the student’s teacher.

I suppose I should emphasize the fact that I am not making the “elitist” point that students are not competent to judge the success of a course as an educational process. The point is rather that some are and some are not and that, consequently, their collective say-so does not constitute a criterion of teaching effectiveness unless we are entitled to assume at the outset that the course was a success ... in which case, why are we evaluating the course and its instructor? My claim is not an ethico-political one of inferiority but a logical claim of circularity.

To this problem, which I admit (in fact, insist) is fairly obvious, there is an obvious partial solution, which might be expressed as follows. While the ideal method of evaluating teaching effectiveness would be, as I have said, to correlate the efforts of a given teacher with the outputs of those efforts, this is would not be possible in this country without fundamental institutional changes. With a few exceptions, such as law schools, professors in the United States evaluate the work of their own students. Of course, it would be implausible to say that we can evaluate the teacher’s effectiveness by looking to whether that teacher judges that the students have learned something from the course, so that the fact that Professor Schmidt gives higher grades than Professor Lopez is evidence that Schmidt’s students learn more. Still, though we apparently are not currently set up to measure teaching effectiveness by monitoring its outputs, we can find proxies for these outputs. By “proxy” I mean a factor that has some suitable causal relation with the output, which is such that this relation enables us to treat the magnitude of the factor as good evidence of magnitude of the output – so good in fact that we may measure the factor as if it were the output. If you cannot observe a fire, seeing the billowing smoke it produces might be just as good. In evaluating teaching excellence, such proxy factors include such matters as clarity and organization. And this is just the sort of thing that student evaluation forms are about. Anonymous student evaluations do enable us to measure professorial clarity and organization, and this is more or less as good as measuring the degree of knowledge that the students derive from the course.

To this I would reply, first of all, that such evaluations are not for the most part used to measure such alleged proxy factors. The evaluation forms, as I have said, typically contain a question that requests a global, all-things-considered rating of the educational excellence of the instructor or the course. This is the question that typically displaces the others in importance, employed by departmental budget committees or by students who view the university’s published evaluation results as in effect the grade that the professor got for that course. This question measures, and is only meant to measure, the professor’s student-approval rating. This is precisely the sort of measurement that raises the epistemic problem of circularity that I have described.

Of course, there are other questions in which the students are asked about specific factors that have some causal relation with the educational aim of the course. But some of these factors raise the same circularity problem that the global question raises. Are the “clarity” and “organization” of the course the kind that conduce to knowledge, or do they represent simplifications and falsifications, in which the richness and diversity of a living discipline are forced into an epistemically arbitrary but user-friendly framework? One is competent to answer such questions only if and to the extent that one has achieved the knowledge that the course aims to impart.

On the other hand, some questions are about factors that do not raise this problem. “Did this course increase or decrease your interest in the subject?” seems causally relevant to the educational objectives of the course, and surely all the students know something about the answer to this one. Further, this question reveals a factor that is causally relevant to the educational objectives of the course. However, I would argue that the answers to such questions cannot be treated as evidentiary proxies for the course objective – ie., knowledge – because their causal relevance to that objective may be either positive or negative. For instance, I often teach courses that deal with subjects – controversial moral issues, film, literary narrative, the philosophy of Nietzsche – in which students often have a considerable prior interest. This prior interest is often a non-academic interest. People often take a course on the aesthetics of film because they have seen movies (who hasn’t?) and enjoyed some of them (who doesn’t?). Some people find that studying film as a rigorous academic discipline actually diminishes the sort of interest they previously had in film. The process of acquiring knowledge includes elements – such as the quest for clarity, the responsibility to justify oneself at every turn, the constant threat of being proved wrong – that tend to diminish certain sorts of enjoyment and interest. Sometimes, a course can diminish interest in a subject precisely because it succeeds in imparting knowledge. (This raises interesting questions about whether knowledge is an unconditional value, but they are irrelevant here.) The effect on the student’s interest in the subject matter is certainly causally relevant to the course objective but, since the relevance can be negative as well as positive, such interest cannot be treated simply as a proxy for achievement of that objective.

In general, the non-global questions on a questionnaire, the ones that seek to identify specific factors relevant to educational success, appear to raise one or the other of these two problems. Either they admit of being causally irrelevant to educational success, so that they raise the same circularity problem that the global question raises, or their causal relevance admits of being either positive or negative. In either case, they seem to be a poor candidates for the role of the means by which we evaluate teaching excellence.

B. The Problem of Academic Standards. So far, I have argued that student evaluations tend to be poor sources of knowledge about teaching effectiveness. Of course mere ineffectuality is not positive harm. As far as that is concerned, the evaluations might produce random results, so that in the long run, after many trials, they have approximately the same effects on everyone. Of course, this is not the case. They are not random at all. We can expect them to be extremely good indicators of one variable: namely student approval of the professor and what the professor is doing. To the extent that they do so, they probably have other effects as well, perhaps both good ones and bad ones. The following is a reason to expect it to have a bad effect, one that should be obvious to anyone who has ever taught a class.

In any department, the faculty will be apt to follow a variety of different grading practices. An assignment that gets a C+ from one professor might well get a B from another. In such a situation, any reason the instructors might have to seek the approval of students will tend to give the instructors who confer the C+ a powerful reason to relax their grading standards, whether they know about the other professors’ grading policies or not. The reason is that they will know something about their own interactions with their students. Some of the students who come to such a professor after being graded by a more lenient colleague will be disappointed by their grade, and will need an explanation for the fact that the grade they have received is surprisingly low. Theoretically, there are any number of possible explanations, but every teacher who has been in this situation knows that there is a substantial tendency to explain the disappointing difference (quite sincerely) in terms of error, prejudice, or gratuitous meanness on the part of the disappointing grader.3 Theoretically, it is possible that the low grader might convince the student that their estimation of the student's performance is "right," or that, whether it is right or wrong, it is based on a legitimate pedagogical approach. But obviously, this is not always going to happen, and some students will think that there is something defective in the judgment or character of the instructor. People in this position will have to either live with a certain heightened level of student disapproval or back off giving the grades that they know are especially likely to cause such problems. There will be a tendency to take the latter course of action, simply because people normally want to be liked and approved of, and they have an aversion to being the object of heartfelt disapproval. If the instructor knows that this disapproval is translated into rating numbers, which then affect the instructor's income, the tendency will obviously be stronger.

The problem of academic standards is, it seems to me, crashingly obvious. It becomes (if that is possible) even more obvious if we consider it in the abstract. What we have here is a process in which the function of participant x is (in part) to continuously evaluate participant y’s performance, where virtually the only means by which we evaluate the quality of x’s performance is to ask y how x is doing. Further, we do this while the process, potentially ego-battering for y, is still going on: we do not even wait until y has had a chance to gain a sense of perspective about what x is doing, and put the matter into context. I cannot think of any other situation in our culture in which we have an evaluation process that fits this abstract profile. At least one reason why we do not do this is, as I have said, obvious: since x’s evaluating y is an essential part of this process, this mode of evaluating x interferes with the process itself. It is important that x’s evaluation of y should track y’s actual performance. Anything that gives x an incentive to evaluate y in ways that track y’s preferences is a distraction at best and a source of corruption at worst. And that of course is exactly what we do when we set up y as the sole evaluator of x.

C. The Problem of the Distortion of Pedagogical Styles. I think I can best introduce this problem by telling a story. Unfortunately, it is a true one. During the sixties4 there was a professor of philosophy, William B. Macomber, whom I knew both as his student and his teaching assistant. Of all the teachers I had, he is the one who had, and continues to have, thirty years later, the most powerful influence on the way I think and the way I teach. He was, it is very safe to say, a controversial teacher. He was most notorious for his Introduction to Philosophy course, which had an enrollment of over 700 students. There was only one assigned reading: one of Plato’s “erotic dialogues” (one semester it was the Symposium and the next time it was the Phaedrus). The exams were all multiple choice, and were meant simply to make sure that students did the readings and attended the lecture. The only other assignment was to write a paper. The lectures were brilliant, but otherwise hard to describe. They were a mixture of argument, epigram, and anecdote. The anecdotes were mainly about his own life. His basic thesis was that the ideals envisioned by the ancient Greeks, especially Plato, have never been surpassed, and that our own civilization is in comparison, denatured and decadent. It has been corrupted in every aspect, but especially in its educational system, by the influence of Christianity. He frequently referred to his own homosexuality, relating it to the homosexuality of Plato, and using the very different attitudes toward homosexuality in Christianity and the Hellenic world to illustrate the (in his view) deep divide between these two civilizations. In their papers, the students were to defend their views on one of the issues touched on in the lectures, and it was expected that in many cases they would of course disagree with the professor.

Like the lectures, student reactions to Macomber are difficult to describe. As I have said, he was controversial: by this I mean that students either loved him or hated him. Someone who is universally loathed is not controversial, no more than one who is universally loved. This of course was no accident. In another of his courses he handed out copies of an essay by the classicist, William Arrowsmith, called “Turbulent Teachers: The Heart of Education,” to justify he own educational practices. In that essay, Arrowsmith argued that the principal aim of a liberal education, especially in the humanities, is to show the student that “a great humanity exists.” Since human consciousness does not normally and naturally have much contact with the ways of thinking represented by the great creators of culture, the function of the teacher must be primarily to go against the grain of our ordingary ways of thinking. Inevitably, this means they must upset us and stir us up. Obviously, this is what Macomber was doing. It was widely believed by the faculty in our department that his courses inspired more people to become philosophy majors than those of any other instructor. Partly for this reason, and also because of his having recently published a distinguished book, some of us were confident he would get tenure. He didn’t, and he never worked in the academy again.

I have often thought of him as an early casualty of the anonymous student course evaluations. At the time Macomber was fired, our department had only been using them for a year or two. All the people who were teaching at that time had developed their pedagogical styles in a completely different regime, in which teaching quality was typically either evaluated by faculty or simply ignored. Some of them were still using methods and approaches that could not well survive in the new system. Those who did not change fast enough would have to face some unpleasant consequences, such as, if one is not already protected by tenure, being fired.

Of course, it would be difficult, after all these years, to show that this is what actually happened.5 However, what is really important for present purposes is to realize that this is just the sort of thing that would happen in a regime of numbers-driven student evaluation of teaching. Arrowsmithian pedagogy is not well adapted to survive in the numbers-dominated academy. The new regime rewards people who can identify, and practice, behavioral strategies that please students. But that is obvious, and it is not the point I wish to make here. The point is that not all strategies of pleasing others are the same, and the new regime systematically discriminates between such strategies. Some of the things that we do that please others are displeasing to no one. They may not please everyone, but they are inoffensive. Others are pleasing to some but displeasing to others. Macomber was a master of the latter sort of strategy. It is entirely the wrong sort of strategy to be using in the numbers-dominated regime. If one student gives me a 5 on the question about my overall effectiveness and another gives me a 1, they do not merely cancel each other out and disappear from the world. They average to a 2.5, which is substantially below average in my department. If I make one student loathe me, I have to get at least one student to love me, just to approach the semblance of mediocrity.

As far as the numbers are concerned, the results of the latter strategy are indistinguishable from those of teachers that the students rate as poor or mediocre. And here we confront a curious difference between the old regime and the new one. Before the numbers-based system of teacher evaluation, a person was regarded as an extraordinary teacher if (but not only if) he or she had a devoted following. The students who reacted with aversion to the same traits that the others found fascinating were more or less invisible. They were simply people who came to one of the teacher’s courses and did not come back. To the extent that the new system holds sway, to the extent that we only look at numbers, what happens is the reverse of this. It is the following that tends to become invisible.

When such a system is linked to pay-raises for teachers, it is obvious that it will result in a massive (if subtle on the micro-level) change in pedagogical behavior. My point is not that this change represents a shift from a superior style of teaching to an inferior style. It is rather that it constitutes an arbitrary narrowing of the array of available styles. Defenders of anonymous student course evaluations sometimes point out that they have virtually done away with a certain obnoxious method of teaching, memorably embodied by John Houseman in the film and television series The Paper Chase, in which the professor motivates students to study by humiliating the ill-prepared in front of the whole class. This, I think, is substantially true.6 I would only point out that it does more than that. It harshly discourages the use of any pedagogical technique that can be expected to be abrasive, annoying, or upsetting to anyone. In the current regime, the most rational course is to choose strategies that are inoffensive.

Surely, this feature of our system of higher education is a flaw, and a serious one. To deny this one would have to claim that all educational methods that are displeasing to a significant part of the student population are, by virtue of this fact alone, bad methods and ought to be discouraged. Such a claim denies the familiar fact of human diversity: different people learn in different ways. To take the extreme case of Kingsfieldian pedagogy: surely there are some students who learn better this way, or such methods would never have existed in the first place. Admittedly, the pre-revolutionary system, in which students were in effect often compelled to submit to teachers of the Kingsfieldian sort, was deficient, but so is to post-revolutionary one, in which they are in effect not allowed to do so even if that is what they want. But this is, as I say, to take what for my argument is merely the worst case. There are many possible styles of pedagogy, and the current regime of teacher evaluation only makes good sense under the implausible assumption that the inoffensive is by virtue of that fact better than the offensive.

Replies to Objections

Such I think are the main considerations that tell against the current system of teaching evaluation. Of course there will be many objections to what I have said. I will conclude by stating and answering some of them.

Student course evaluations are reliable indicators of at least one factor that is always relevant to success in teaching, and it is one that students clearly do know about. This is a response to my argument in (A). In an article offering a limited defense of student evaluation of teaching, philosophy Professor Kenton Machina asserts (emphasis in the original): “Student evaluations (if honestly conducted) basically report the extent to which the students have been reached.7 Machina does not tell us what this “being reached” consists in, though he does add to this assertion by saying that the evaluations measure whether students have been “reached educationally” (emphasis added). I would like to know whether this state, of being reached educationally, is it an emotional state or a cognitive state: does it consist in feeling something or in believing something? And what feeling, or what belief, does it consist in? Machina also fails give any evidence for thinking that this state or condition, whatever it might be, is actually measured by evaluation forms.

Nonetheless, the mere fact that he would state the idea so flatly, without justification or explanation, indicates that it must be extremely plausible to some people and, if only for that reason, it deserves to be investigated. Some light on what it means, or might mean, is shed by a recent book, What the Best College Teachers Do, by Ken Bain.8 This book reports the results of a fifteen year long project in which Bain studied the almost one hundred of the “best” teachers at colleges around the U. S. In explaining how he decided which teachers were the best, he says he had two sorts of criteria. He considered such things as how well the teacher’s students did on department-wide exams, numbers of students who went on to achievements of their own, and interviews with former students on how well the teacher’s course had prepared them for subsequent activities. Obviously, these are examples of the “ideal” sort of evaluation I discussed above: evaluation by results. Prof. Bain’s account is very sketchy on the matter of which of these sorts of tests he applied and to how many of the instructors he applied them.

Concerning the other sort of criterion, however, he is much more clear: to be regarded as one of the “best” teachers, the subjects would have to be rated very highly by students. His defense of taking high student approval as a necessary condition of teaching success is clearly influenced by Machina’s article (which Bain cites). I will need to quote this defense virtually in toto, to assure the reader that I am not misreporting it:

We wanted indications from the students that the teacher had “reached them” intellectually and educationally, and left them wanting more. We rejected the standards of a former dean who used to say, ‘I don’t care if the students liked the class or not as long as they performed on the final.’ We too were concerned with how students performed on the final, but we had to weigh the growing body of evidence that students can ‘perform’ on many types of examinations without changing their understanding or the way they subsequently think, act, or feel. We were equally concerned with how they performed after the final. We were convinced that if students emerged from the class hating the experience, they were less likely to continue learning, or even to retain what they had supposedly gained from the class.9

This passage contains two dizzying logical leaps. One is the inference from the fact that students like a course and approve of the instructor to the conclusion that they must thereby acquire some of the love of the subject of the course, a love which results in further inquiry. Surely there are any number of reasons for enjoying a course and approving its instructor which are perfectly compatible with a disinclination to pursue the subject further. The other inference starts with the plausible claim that if students hate a course they will be less likely to pursue the subject in the future. Here it is more difficult to say what the conclusion is supposed to be, though to be relevant to the issue at hand it would have to be something to the effect that the more students like a course the more likely they are to pursue the subject further.

The notion that seems to lie behind the idea that evaluations measure whether students have been “reached” educationally seems to be that, since real learning is not merely a cognitive process but an affective one as well, then even if the forms only measure students’ emotional reaction to a course, they still measure something that is causally related to the indented outcome, which is knowledge. This is undoubtedly true, but as I said before, it makes all the difference what exactly this causal relationship is. Surely the most plausible way to conceive it is to suppose that some degree of enjoyment is a necessary condition of learning. This is clearly the view of Prof. Machina, who explains the italicized sentence I have quoted by saying: “Very well-organized lectures, ... and the like may all be positive factors of some kind, but if they do not result in reaching the students, they will not constitute effective teaching.”10 This is the sort of relationship that might obtain if, to invent an analogous case, human beings were unable to swallow a pill unless it were sweetened by being coated with sugar. In that case the sweetness of a pill would be a necessary condition of its medical effectiveness. But it would not follow that sweetness is a proxy for medical effectiveness, because the causal connection is of the wrong sort. It does not support us in thinking that the sweeter it is, the better it is: that the best pills are the ones that are most highly rated for sweetness. Similarly, there might be some level of positive affect in an educational experience that is high enough for the creation of knowledge to occur. It could also be true that in some subjects the highest levels of student approval are signs that something educationally corrupt is going on, that the pill has been loaded with medically deleterious empty calories.11

The empirical evidence does not clearly show that student course evaluations lower academic standards. This of course is a response to problem (B). Machina argues that, though there is a statistical correlation between giving high grades and receiving high evaluation numbers, that does not prove that the high grades cause the high evaluations. They may have a common cause: students learning more.

If the data show a positive correlation between student opinion of a course and grades received in the course, it nevertheless remains reasonable to suppose those students who received good grades learned more, and as a result viewed the course more positively.12

Surely, it is not reasonable to suppose this at all. Such a supposition flies in the face of a fact that every student knows: that sometimes the reason why the students in one class got higher grades than those in another is that their teacher is an easier grader.

On a deeper level, Machina does have a sound methodological point: the fact that two variables are correlated does not prove which one causes the other, or whether there is some more complicated causal relationship between them. Once a correlation has been noticed, the question is: what is the most plausible, best-grounded explanation of the correlation? In (B) I explained the correlation by supposing that students can notice discrepancies in grading and that they have a tendency, to some significant degree (which may be far from universal), to attribute lower grades to teacher error. It is important to notice that my explanation does not deny the existence of the sort of causation that Machina’s explanation describes. On the other hand, his hypothesis is that the sort of causation I describe does not occur: high grades do not cause any high evaluations. Rather, he is asking us to suppose that all of the correlation between high grades and high evaluations is due to a) teachers accurately reflecting student learning in their grading practices (no arbitrary discrepancies from one teacher to another) and, in addition, b) students accurately reflecting how much they learned in their questionnaire-filling practices (and not being influenced by their grades). Which of these two explanations, his or mine, is more plausible?

The use of questionnaires to evaluate performance is much more common than I make it sound in (B). Here the objection would be that something analogous to student evaluation of teaching is in fact very common in our culture: companies use consumer questionnaires to evaluate new products, politicians test their performance by using public opinion polls, television shows are renewed or canceled based on Nielsen ratings. How is basing university personnel decisions on student ratings any different? My answer is that it is different in at least two ways. First, in none of the above examples is it part of x’s function to evaluate the performance of the y who is filling out the questionnaire. Thus none of them involve the unique potential for distraction and corruption that I have pointed out in the case of teaching evaluation. Second, none of the above examples are cases of evaluating the intrinsic excellence of x’s performance, but rather its potential for success in the market place. The people who use these questionnaires are trying to find out how many consumers will buy their product, how many voters will vote for them in the next election, and how many viewers will see their sponsors’ commercials. Evaluation of teaching is (at least officially) intended to measure the quality of the teaching itself, and not how many consumers (ie., students) it will please.13

The pedagogical styles that are encouraged by student course evaluations are ipso facto superior to those that are discouraged by them: they are morally superior. This objection, a direct response to (C) above, is one that I have never heard stated, but I suspect that it is at bottom the real reason why the regime of student evaluations continues to exist virtually unchallenged. After all if, as I have said, problems (a) and (b), which raise the question of whether it is rational to measure and evaluate professorial conduct in this way, are obvious, it is a real question why we continue to do it. It is a paradox that calls for a resolution. Having posed it, though, at least part of the resolution is readily apparent: student evaluations are not meant to measure and evaluate professorial behavior at all: they are meant to shape that behavior. The objection is that only my argument in (c) goes to the heart of the matter, but that it gets the matter precisely wrong. The behaviors that are penalized by teaching evaluations are all bad ways to treat people. They are bad because they are morally bad. Generally, the good ways of treating people are things that they will like, and the bad ways are things that they will not like. This is the idea behind the golden rule: do unto others as you, so to speak, would like them to do unto you. The sort of pedagogy that comes out of this ethic is a nurturing pedagogy, one that supports student self-esteem, and kindly encourages their efforts to learn. For the same reason, it is also the sort of pedagogy that is encouraged by the student evaluation system.

It may well be that this kindly ethic of warmth and nurturance will, as Nietzsche claimed, tend to dominate people who live in a democracy.14 Notwithstanding that, it is subject to some very serious objections. As I have characterized it, it is a species of utilitarianism, in as much as it characterizes the ethically good as consisting in giving people what they want. Utilitarianism is per se a perfectly respectable ethical doctrine. However, this is a very peculiar sort of utilitarianism, in that it conceives the good as consisting in giving people what they want now, at the time that it is given to them. It is, we might say, short term utilitarianism. Every respectable variety of the utilitarian doctrine is based on long term utility. According to long term utilitarianism, the familiar cliché of parenting – “You’ll thank me for doing this in ten (or five or twenty) years” – should (whenever it is true) win every argument. Most utilitarians would probably agree that the ultimate standard for deciding the duty the teacher has toward his or her students is the students’ preferences – but they would maintain that the genuine moral standard is the satisfaction of student preferences in the long term. It certainly is not the satisfaction of the preferences they have when they are eighteen years old, and while the course is still in progress. It is certainly very tempting to treat others in accord with their current preferences, but whenever these preferences conflict with their most likely long-term interests they are, according to the sounder sort of utilitarianism, only a temptation: it is our duty to act contrary to them.

Such words as these do leave rather a bad taste in the mouth, I admit: they seem paternalistic. They require us to see students’ present selves as incomplete and deficient and so they seem to speak of them with disrespect. But consider for a moment what the alternative is: To treat people’s presently occurring desires as the ultimate standard of value, and to treat offending others as the cardinal sin. This is to deny them the possibility of growth, and to deny the openness to pain and disappointment which their growth requires.

1. The sort of thing I am most curious to know for purposes of improving a course include such things as which reading assignments the students liked the most, or the least, or which exercises seemed pointless to them. For good reasons (some of which are obvious) personnel-oriented evaluations never ask such questions. Personally, my advice to teachers who want to use evaluation forms to decide how they may improve their educational approach is: design your own forms, with questions designed to provide the sort of information you wish to have, and distribute them in addition to (or even, if permitted, instead of) the administratively dictated forms.

2. Roughly this system, I am told, is still commonplace in Europe. See Harry Brighouse’s discussion in his contribution to this volume. “Grade Inflation and Grade Variation: What’s All the Fuss About.”

3. A widely accepted theory in social psychology ("self-serving attribution error") holds that people tend to attribute their successes to themselves ("I got an A!") and their failures to others ("She gave me a C!"). See Miller, D. T., & Ross, M. (1975). "Self-serving biases in the attribution of causality: Fact or fiction?" Psychological Bulletin, 82, 213-225.

4. As many people have pointed out, when we say “the ‘sixties” we generally are referring to the years 1965-1975. In this case the years involved were 1969-1972.

5. For whatever this information might be worth, I recently asked him about the evaluations he got in those courses, and he said that all he could remember was that “they were dreadful,” and that they were noticed by the people who had control over his tenure decision.

6. According to a legend, which can still be found on the internet, when the book on which the film was based first appeared, several Harvard law professors boasted that the Kingsfield character was based on them. Today, it is almost impossible to imagine someone bragging about a thing like that – even at Harvard.

7. Kenton Machina, “Evaluating Student Evaluations,” Academe 73 (May-June 1987), pp. 19-22. The passage quoted is on p. 19.

8. Ken Bain, What the Best College Teachers Do (Cambridge, Massachusetts: Harvard University Press, 2004).

9. What the Best College Teachers Do, p. 7.

10. “Evaluating Student Evaluations,” p. 19. I should point out that Machina’s defense of evaluations is very limited, and he in fact argues against the crude, numbers-driven reliance on them that is defended by Bain. He says, for instance, that it would be a “serious error” to use student evaluations to judge the intellectual quality of a professor’s lectures.

11. This is probably true of some branches of philosophy. Philosophy often takes the very ideas we live by – God, freedom, immortality – and subjects them to ruthless logical criticism. If it is done right, it is very disturbing, a sort of gambling with one’s way of life. Some amount of negative affect is inevitable, unless the professor is falsifying the subject by cushioning students against the shocks – and there are many, tempting ways of doing that. No doubt, similar considerations would apply to a number of other humanistic subjects as well, in addition to philosophy.

12. “Evaluating Student Evaluations,” p. 20.

13. Here is one way of appreciating the unique status of anonymous teacher evaluations in our society. They are the only example in our socio-political system of power that is routinely exercised in the complete absence of responsibility. Politicians are subjected to checks and balances, judges are subject to judicial review, business people are subject to regulation and to the discipline of competitive markets, witnesses in court are cross-examined, but because the evaluations are anonymous there are never any consequences to filling them out in an irresponsible way.

14. This at any rate is how I interpret Nietzsche’s celebrated discussion of “the last men” in Zarathustra. See The Portable Nietzsche, trans. and ed. by Walter Kaufmann (New York: 1976, Penguin Books), pp. 128-131.