I remember the dead silence in a faculty meeting when a Notre Dame Dean said over 60 percent of the grades in English and Humanities classes were A’s, the rest were A-minuses and B-pluses. It was an attempt to embarrass professors’ lack of effort at identifying variations in student performance. Shaming didn’t work, nothing. And, as a new report shows, Notre Dame is not alone. Grade inflation is alive and well.
Stuart Rojstaczer has just updated his scathing analysis of the increase in grade inflation.
He may be the messenger that elite colleges and their professors will want to kill but he has done yeoman’s work exposing the industry practices. Using data on grades from over 210 U.S. colleges and universities, and over two million undergraduate students, Rojstaczer has found that grade inflation is increasing and private colleges are the worst offenders.
Perhaps the increase in overall GPA’s is because students as a whole are getting smarter? Not according to this report—it does NOT find that grade inflation can be explained by an increase in the quality of students. Harvard, Yale, and MIT have the worst grade inflation, as one of his charts shows. The higher the school is above the blue regression line, the worse their grade inflation. (There are some qualifications on these data, but you’ll have to read the report.)
Other variances are pretty interesting as well. For example, colleges that focus on the sciences have had less grade inflation than those that do not. And some schools have successfully focused on fighting the trend. Reed, Princeton, Columbia, and Wellesley have explicit policies to hold down grade inflation, and Rojstaczer’s data shows those policies seem to work.
Rojstaczer’s methods are debated but I suspect that most professors know in their hearts that he is on to something. High grades make happy students, and happy students don’t complain. After many years on the peer-review committees awarding tenure and promotion, I always reported student evaluations results and the average grade. In my sample, the higher the average grade the more pleased the student was with the teacher.
I am not alone. Michael Huemer writes, “Despite some dissenting voices, the influence of grades on student evaluations seems to be an open secret in colleges and universities. In one survey, 70 percent of students admitted that their rating of an instructor was influenced by the grade they expected to get. Similar proportions of professors believe that grading leniency and course difficulty bias student ratings.”
There could be good reasons for the correlation: Functioning classes produce pleased results. The less attractive reason is that the teacher inflated grades, making more students satisfied with the course, and the professor.
So grades are inflated, teaching loads have fallen, professor and college administrative salaries are up, tuition increases have beaten all inflation records, and more and more classes are taught by adjuncts or graduate students. Inflation all around. But remember the economist’s definition of inflation—an increase driven by excessive demand, but with no fundamental improvement in underlying quality.
4 Responses to Prizes for All: Grade Inflation Is Alive and Well
luther_blissett - April 30, 2010 at 12:00 pm
Grading at the college level is a joke all around. Even these “tough on inflation” schools have no sane grading policies, no means of adjudicating whether an A in Biology is akin to an A in Economics, whether the workloads of various classes across various majors at various levels are similar, etc. And from school to school, there’s no way of comparing A students, B students, and so on. An “A” in American Lit I at Yale means nothing more, at this point, than that a single professor at a particular point in time gave a particular student an A.And then there’s the problem of schools with no set curriculum that grade students on prior skills without teaching them. So a student takes a history course, writes an essay, gets a C because it was poorly organized and had no thesis — but the history professor never taught his class how to organize an essay or what constitutes a thesis in history. So students are graded on what they learned from other classes, not from the particular professor doing the grading.
intered - May 2, 2010 at 2:43 am
I’m not certain that anything meaningful position can be developed on this topic in 500 words. It is possible to identify a few important subtopics that have been omitted.First is the large distinction between criterion and norm-referenced grading. The former assigns grades based on evidence that specified learning outcomes (knowledge, proficiencies, etc.) have been achieved to a specified criterion level while the latter assigns grades based on relative performance. Based on this distinction, even in its simplest form as I have expressed it here, no one should be surprised to learn year/year grade distributions are relatively uniform in the sciences and therefore that science grades have not inflated as much as some other disciplines. Performance evaluations in the sciences are most often criterion referenced, whereas performance evaluations in courses where intersubjectivity is more likely to be a part of the assessment, perhaps even essential to good instruction, tend to reflect norming in the assessment process. This norming might be intentional or an unconscious effect of the tendency most of us have to adjust our teaching and evaluations standard to a level suitable to the particular group of students in front of us. Second, some disciplines (again the sciences are an example) assess to a relatively large number of knowledge components demonstrating high interrater reliability. Ten physics instructors will all agree that the student did or did not correctly label the particle. Interrater reliability is low, often very low in the so-called “softer” disciplines (non-pejorative) thus year/year grade distributions will vary more widely and inflationary forces, whatever their origin, have a better chance to exert themselves. The low interrater reliability scores on most essay evaluations, even those employing reasonably detailed rubrics, offer good examples.Third, one can legislate lower mean GPA with institutional policy but I would think that an economic policy analyst would speculate as to the downstream effects of such a policy. What I have seen happen in these situations cannot be called better grading behavior. Knowing that their mean GPA is being examined at an administrative level, instructors tend to lower their mean in the most statistically savvy way, by focusing on situations where an ‘F’ or ‘D’ can be awarded. When you compare the middle and upper part of the instructor’s distribution pre/post policy the distributions are relatively unchanged. A sound way to bring grading behavior into line is to focus on variance, the higher the better within reason. The variance statistic can’t be gamed like GPA. Assuming students will not sit still for unfair or capricious grading, high variance means that the instructor is making defensible distinctions among individual levels of performance and attaching grades to those distinctions. I believe this is an essential component of responsible teaching. I see this issue, by the way, as a contributor to the grade inflation problem, such as it is.Fourth, I say “such as it is” because analyses of grade inflation commit fallacies of misplaced precision, confidence, validity, and generalizability (the latter two being different ways of expressing the same thing). For most courses taught at most institutions, there is no way to create a grading system in which grades are commensurable from one institution to another, or even one department to another. ‘luther_blisset’ makes this point above. There are many reasons why this is true. What grade will ‘A’ level performance in an Econ course at BigStateU earn you at Reed? An ‘A’ is probably not a good guess is it. Material threats to reliability, validity (most instructors’ tests earn low grades in a measurement science), commensurability, and more limit what there is to mean by a grade when you attempt to export the material fact very far beyond that specific classroom.Fifth, it would seem that the concept of ‘value added’ should have a voice in this discussion. A Reed student gets an ‘A’ on a test that he would have scored an ‘A’ on had he not taken the course. A student at BigStateU gets a ‘C’ in the same course where the grade represents considerable work on his part and a 250% increase in his knowledge on the subject. This is where norm referencing often comes into play when we figure out how to teach and grade the students in front of us. Should it? I don’t know. Sixth, my work in this area leads me to conclude that the most tractable cause of grade inflation is too few high stakes metrics taken too infrequently with tests of low validity, the classic example being a five credit course where the entire grade is determined by a mid-term and a final. Instructors who take dozens of small measures of different learning attributes across the span of the course tend to have high variance statistics (this entails lower GPAs). They also have students for whom the final grade is not a surprise, even if it is low, because 50% of their grade was determined at the 50% mark, 75% at the 75% mark, etc. There are social pressure issues as well but the entire discussion requires a different forum. Last, I’m disappointed to see the tired apocrypha linking grades and student evaluations dragged out again when it is so unscientific and unprofessional to do so. Having run dozens of these regressions over the past 25 years, the largest being 85,000 records, I can tell you that your students, in the aggregate, are probably neither so shallow nor so unintelligent as to barter evaluations for grades, perceived or otherwise. Set aside for a moment the fact that some instructors have so little confidence in their students. What in the world are you saying about the professoriate by suggesting that they are trading their integrity for a better chance at tenure in sufficient numbers as to cause grade inflation? We should do something about grade inflation but it is not a crisis. The only thing that makes it seem so is misplaced confidence as to what there is to mean by “grading behavior.” If you want to curtail grade inflation: (a) teach instructors how and how often to evaluate, (b) begin measuring variance, and (c) make this information available to serve as constructive feedback. Policy change is probably not necessary.Robert W Tucker
arrive2__net - May 3, 2010 at 1:23 pm
A professor at Dartmouth noted in a Chronicle article that when Dartmouth upheld high grading standards, their deserving students had trouble getting into grad school. (http://chronicle.com/article/How-Students-at-Dartmouth-C/21586/). Loyola Law School Los Angeles recently announced giving whole cohorts of graduates a GPA boost because strict grading standards were preventing many of them from getting desirable jobs (http://chronicle.com/article/A-California-Law-School-Wil/64949/). Of course there are scholarships out there that require the student to have a certain GPA, so if the school wants to retain students with that scholarship they had better be willing to have some flexibility with grades. My point is that there are a lot of pressures for grade inflation that go beyond the institution’s control. Bernard SchusterArrive2.net
goxewu - May 11, 2010 at 9:02 am
Re #3:”There are a lot of pressures for grade inflation that go beyond the institution’s control.”Yeah, the pressures are beyond the institution’s control but, one hopes, whether to cave into them or not is not.What’s the uncrossable line between giving students inflated grades* so they can get into graduate school or get scholarships, and giving them inflated grades so that they can graduate this semester, or maintain a perfect 4.0, or stay eligible for lacrosse, or not have their parents yell at them? I don’t think there is one. This is a case of a slippery slope that really is a slippery slope.Just what society needs, too: More mediocre students (this time, Dartmouth’s on top of everyone else’s) getting into law school they shouldn’t get into and becoming the lawyers they shouldn’t become. Thank goodness Mr. Schuster isn’t talking medical school.* A = excellent, B = good, C = average, D = poor, F = fail. Why do so many academics have a problem with that?