wp home > junior year writing > sourcebook for instructors > high stakes and low stakes evaluation

High Stakes and Low Stakes Evaluation of Writing

By Peter Elbow

In this essay I suggest two ways to make the grading of writing easier, fairer, and more helpful for students: the use of minimal grades or fewer levels of quality--not only for low stakes writing but also for high stakes writing; and the use of criteria in which we spell out the features of good writing that we are especially looking for in any given assignment.

No matter how well we write comments in words, in the end we are always stuck with the quantitative question: "But what grade did I get?" For students, the grade usually overshadows the comment. Even teachers tend to use the phrase "grade a set of papers," when really they will spend far more time on commenting than grading. I have always known that students were anxious about grades. I had not realized until recently how many teachers were anxious and full of misgivings about grades. I am relieved to discover I'm not the only one.

Till now I have tended to feel that the problem with grades is their quantitative or numerical nature. But my misgivings and perplexity have finally led me to roll up my sleeves and dive into the dangerous element itself and focus on exactly that dimension. My point in this essay then is that we need to think harder about grades in their nitty-gritty quantitativeness--to be more analytical, even algebraic about them. If my approach seems "hard" and literally calculating, I think you will see that my goals are nevertheless "soft": a more generative, supportive, and less adversarial atmosphere for learning and teaching.

High Stakes and Low Stakes Grading

I am assuming here (like most teachers, surely) that we need to demand high stakes writing in all college courses and take it seriously. We can't tell whether students really understand what we are teaching unless we see their writing (or have extended conferences with them), so their writing must play a significant part in determining their final grade. But then comes the grading of high stakes writing! It's so difficult. It causes such doubts and worries. We feel so bad when we think about students who may know the material but can't demonstrate it in writing because of their anxiety or lack of skill. And what about the ones who somehow "show" an undeniable understanding, yet their writing is undeniably awful? And in the back of our minds may lurk misgivings about our own writing skills. And how much do we grade down for awful writing? For lack of clarity? For mistakes in spelling and grammar? And since high stakes writing counts for a lot, students often argue with us about our grading of it--which is doubly hard to take when we don't feel so secure in making these fine-grained distinctions with pluses and minuses up and down the scale from A to F. In short, high stakes grading makes for perplexity and anxiety.

What can we do about it? Let's do what we usually do when things get hard in grading. Let's stop and get up and walk around and do something else. Maybe it'll be easier when we come back.

That, in fact, is my serious strategy here. I think I will be able to offer some help on high stakes grading if I drop it and circle around for a while and then come back.

Let's concentrate for a while on low stakes writing. The widespread use of low stakes writing in writing intensive courses has led to what I would call low stakes grading. Such grading is often the result of certain pragmatic pressures. That is, faculty who use writing for learning are often not writing teachers and they are often dealing with large classes that are not defined as writing courses. Thus, these faculty often feel they cannot grade the writing carefully or elaborately--and they don't feel obliged to do so since the pieces of writing don't count for so much. In this situation teachers have used a variety of what might seem like unconventional ways of grading:

The most obvious advantage of this low stakes grading--I call it "minimal grading"--is pragmatic: it makes life easier for us. It's much less onerous to read student writing when the grade is quick and easy to give and we don't have to comment on it. But there are also advantages for student learning.

A low stakes or minimal grading scheme may not give students an incentive to strive for an A for excellence (though we get a bit of this incentive if we use a three-level scale). But the tradeoff is that we get to ask students to write far more than if we had to grade everything carefully. We get to ask them to think actively about far more of the course material. They have to answer the questions, get their thoughts into writing--yet they don't have worry so much about whether they are getting it just right, or writing in the way that the teacher likes, or saying what the teacher agrees with.

Motivation. Grading is often defended as a source of motivation. But when students struggle for excellence only for the sake of a grade, what we see is not motivation but the atrophy of motivation: the gradual decline of the ability to work or think or wonder under one's own steam. Low stakes grading, however, helps students gradually develop a bit of genuine or intrinsic motivation--to develop some of their own curiosity and standards. They get a time out from their habitual and understandable preoccupation with "What is my teacher looking for? What are my teachers' standards?" but rather, "What do I think? What am I looking for? What is interesting to me? What are my standards?" Of course students nurtured in a grading economy often need some extrinsic motivation to get them to work. But that's exactly what low stakes grading provides. Students have to do the writing and engage the material, but they have a lot of choice about how to do it. This provides small protected spaces for gradually developing small bits of intrinsic motivation. And of course they still have some high stakes assignments graded in a high stakes way--assignments where we provide most of the motivation. (I've written more in the introduction to this volume about the advantages of low stakes writing.)

Before going on, let me add an important point of strategy about introducing low stakes grading at the start of a course. For the first few low stakes pieces, I find it helpful to use a two-level scheme such as pass/fail or satisfactory/no credit. That is, I seek a middle ground between using three levels (with its possibility for distinction), and not grading at all (where I accept anything they turn in or don't even collect their writing). If I start with a three point grading scale, too many students become too preoccupied with getting that "strong" or "excellent" and thus with trying to figure out what I "want." They fall back into writing for the grade instead of writing for learning. If on the other hand I accept anything they write or don't even collecting their writing, I fall into the opposite problem: I exert no pressure at all on students to engage the issues and thereby tempt students who have never done this kind of writing not to try at all.

My strategic decision to start with two-level grading illustrates my overall goal in using low stakes grading. I am trying to teach students an important skill that is surprisingly hard for many of them: the ability to write about a topic with their mind focused entirely on the issues, and not on how their writing will be judged by a grader. Almost all students have a hard time focusing their attention wholeheartedly on an academic topic because they so naturally associate writing with high stakes grading, and thus are distracted by questions of how the teacher will grade ("What is s/he looking for?") and with questions of style and quality. As a result they are unable to focus all their attention on thinking and exploring a topic as they write. Consequently they have not experienced the satisfaction or even pleasure of writing seriously about an academic topic.

After they have done three to five pieces of low stakes writing graded on a two-level scheme--that is, with a great deal of leeway but still a threat of no credit for goofing off--I can move my low stakes grading in two directions if I want. I can start doing some low stakes writing in class that I don't even collect--and trust that they will make good academic use of the time. And I can start also using a three level scale and trying to pushing them harder with the reward of excellence--without it just leading to clenching and hyper-attention to "what I want" and "style." In effect I can begin to have a spectrum of gradations from low to high stakes writing. But in fact I often stick with the simplicity of two-level grading.

A few students are confused at first by this dialectic between low stakes and high stakes grading. Despite my explanations and warnings, they are caught off balance the first time I use high stakes grading on a high stakes essay. They assume that because I started off with low stakes grading for a number of assignments, I was a "low stakes kind of a guy." They are used to pigeon holing teachers as either high or low stakes--either hard or soft. I find I have to be extremely explicit and even repetitive to help them realize that this whole approach enacts a dialectic. Indeed I like to bring in the word "dialectic" and talk about the psychological and intellectual benefits of unresolved contraries. But sometimes I do better with the bluntness of saying, "Perhaps you better think of my grading policy as schizophrenic."

Untangling Two Major Variables in Grading: Stakes and Levels

We tend to associate low stakes grading with the use of levels such as pass/fail. Thus if we want to raise the stakes and want students to strive harder, we assume we should grade on more levels: conventional A through F grades feel more serious and make the writing seem to matter more.

But these are mistaken assumptions. We can grade low stakes assignments with multiple levels; plenty of teachers use conventional A-through-F grades on low stakes pieces. And we can grade high stakes writing assignments with few levels. For example, at MIT for the last twenty years or so, faculty have given nothing but the final grades of "pass" and "fail" to all first year students in all their courses. The stakes are very high indeed and so are the standards, but only two levels are used.

The important strategic point here is that stakes and levels are quite different. We will have a much better time grading high stakes assignments if we untangle these two quite separate variables or dimensions of grading--two different questions:

It is indeed true that if we add more levels we add more opportunity for the recording of excellence: more opportunity for students to distinguish themselves, to excel, to surpass others or be surpassed--more competition, more pecking-order. But giving students a chance to have excellence noted and recorded is not enough to make all students strive. The chance to get an A rather than a Pass will only make students work harder if a number of other conditions are satisfied:

In short, increasing the number of levels in our grading scale is an uncertain engine for making students work harder. But it is a certain engine for making us work harder. That is, more levels mean more decisions to make--decisions which are by the same token harder to make because they are more fine-grained and therefore more debatable. And by making more fine-grained decisions, we are by the same token providing students with more occasions to dispute and resent these very decisions that were hard to make ("What do you mean B-minus? That was a solid B!").

But if instead of assuming that we have to use a conventional A-through-F scale for any high stakes essay, we think more strategically about the use of both variables in grading--stakes and levels--we can make students strive harder for excellence yet not increase our work much or at all, and not create so many opportunities for students to resent or quarrel with our grade. This will help keep down the adversarial spirit around grading.

Thus we need to consider carefully the range of options we actually have at our disposal in grading high stakes assignments:

I'm suggesting then that when we have to grade high stakes essays, we put the emphasis on the fact that we have raised the stakes--and realize that we don't have to use the multiple levels of conventional grades. These pragmatic considerations are particularly important if we are teaching a large class. If we are looking at a big stack of papers, consider the difference between giving each one a conventional grade from A to F versus just having to pick out the ones that are notably strong or notably weak.

I imagine two quarrels with my line of thought. (1) "What you are suggesting isn't much different--or easier--than what we do now. Most of us only use a few out of the ten or so possible levels." But it is exactly this restricted use of the grading scale that has exacerbated two problems we see all around us: grade inflation and the fact that no one knows what conventional letter grades mean. A general principle from linguistics states what we already know: the more widely used a symbol is, the more widely divergent and ambiguous its meanings will be. B, for example, has become an almost empty term. In some teachers' hands it means "good strong competent work"--after all, B is an honors grade in most colleges and universities; in the hands of others it means "somewhat disappointing, second order work"; and for many students it means "unsatisfactory." (And even a restricted range of grades from A-minus to C-plus means having to choose among five levels.)

(2) "It sounds childish or amateur to use grade school words like weak, strong, poor, satisfactory, excellent. It isn't fitting for higher education." But I find these homely words a vast improvement because of their clearer meanings--instead of standard grades that lead to such misinterpretation.

Besides (and I have left untill the end what is probably the biggest argument), why go to the trouble of trying to distinguish between a B and a B-minus--much less a B and a C--giving ourselves more work and our students more occasions for argument--when the resulting grades lack not just meaning but fairness. Careful research has demonstrated over and over again what common sense has told us and what our students have learned through experiments with submitting the same paper to different teachers: good teachers and evaluators routinely disagree widely with each other over levels of more than a full grade. Fine-grained evaluations of writing cannot be trusted. If we use only two or three levels, we vastly reduce the untrustworthiness simply by sorting performances into fewer boxes.

"How Can I Calculate a Regular Final Grade if I Have Nothing But Minimal Grades to Work With?"

This problem brings up a third important variable in grading: the number of assignments we grade. It is certainly true that if we only have two or three graded assignments, and they are graded on only two or three levels, then we are in trouble when we try to calculate the final grade for the course. But if we have lots of grades--which is easy with lots of low stakes assignments-then the final grade is no problem.

Take an extreme example. Suppose we have a course with weekly low-stakes writing assignments judged on a two level scale, two high-stakes essays judged on a three level scale, and a conventional final exam judged on three or four point scale. (It's not so hard or unreliable to use four levels for an exam if the exam contains multiple answers.) One easy way to calculate the final course grade here would be as follows: students who have a "satisfactory" on all the low-stakes pieces start off with a foundation of B. Then the two high stakes essays and the exam determine whether their B gets raised or lowered. In effect, students who keep up satisfactory work on the extensive weekly reflections on the course material start off with the presumption of a B and are pulled up if they manage to distinguish themselves on the high stakes pieces--or pulled down if they do badly. Students who have some unsatisfactory low-stakes pieces--who don't keep up with the weekly reflections on course material--start off from a lower foundation, and they are similarly pulled up or down by their high stakes work.

This formula gives quite a lot of emphasis to the low stakes assignments: individually, their weight is low, but in sum they count for a lot. I want my students to take these pieces seriously--that is, to work hard but not in a worrying way. Not everyone will agree with my priorities here, but there are obviously other formulas one could use for this situation--more sophisticated ones. I just want to illustrate in the simplest way my point for this section: we can easily derive a maximum-level final grade from minimum-level constituent grades. (Notice I didn't even talk about other factors which many of us feel are important ingredients in calculating a final grade: attendance, participation, effort, and improvement.) In short, using more writing assignments doesn't just help students learn the material, it also makes it easier for us to compute final grades.

Portfolios. Portfolios can be a big help in calculating a conventional final grade while still using minimal grading for individual performances. A portfolio system makes it natural to use minimal grades even on high stakes essays because those initial high-stakes grades are, in a sense, conditional. The essays will show up again in the portfolio at the end when it comes time to calculate the final course grade--this time using the conventional scale of grades. Portfolios thus permit us to wait untill we have a rich collection of many pieces of writing before having to make our final judgment on the scale from A to F.

Portfolios are particularly useful for enhancing learning. They invite us to get students to look back over all their work for the course and reflect on what they have learned--and even to analyze their learning process--because the most important piece in a portfolio is the reflective essay about the contents of the portfolio. Low stakes assignments create much more learning if you ask students to gather them all together at the end and articulate the most important insights they have gained--both about the course material and about their processes of learning, thinking, and writing. (See Fulwiler's suggestion in this volume for a portfolio of letters written throughout the course.) Low stakes writing provides students a good window for reflecting on their own habits of thinking and using language. Readers who want to open the door into the rich literature on portfolios can start with the Belanoff and Yancey collections. I love the way portfolios permit me to invite students to put some of their low stakes writing into the final high stakes pot--a satisfying thing to do since some of their low stakes writing is often very good.

Portfolios are probably not so feasible for very large classes. You can't even carry one hundred portfolios back to your office. But in a large class, it is perfectly feasible to ask students to write that final paper that articulates their most important learning and reflects on or analyzes all the writing they have done in the course. Many teachers find it helpful to ask students to suggest their final grade in this final essay of self-analysis.

Contracts for grading. I tend to use a grading contract ("If you do x, y, and z, you will get such and such a grade."). A contract makes it downright easy to come up with a conventional final grade--even though I give only the most minimal grades on their writing over the course of the semester. I think I can create more learning by trying to figure out what activities and behaviors will lead to learning and specifying them in a contract than by trying to make dubious, fine-grained judgments about the quality of student work. If I can make a good contract by figuring out things to ask for that get students to throw themselves into their work, I know I have promoted a lot of learning. On the other hand, even if I do a "good job" in giving conventional grades to a set of papers (and I'm not sure any more what that means), I'm not sure I have promoted much learning.

Notice, by the way, that conventional grading is an indirect way to try to get students to work hard: we hope they will work when we hold out rewards or punishments for the product of their work. But this doesn't always work; there are quite a few students that our grading doesn't coax hard work out of. A contract permits me to be direct about it and simply specify openly the work or behaviors that I think are important. (See the appendix for a sample contract.)

Criteria for Grading

When students ask the inevitable question, "What are you looking for on this paper?" I want to say, "Oh I don't know. Just surprise me with something wonderful." When students ask us why they got an unsatisfactory, I want to say (with Louis Armstrong), "If you've got to ask, you'll never know." But those answers don't work too well. If we grade, we have to try to articulate publicly the criteria behind our grades.

The criteria for low stakes assignments don't usually cause problems. But still we need some clarity. For low stakes assignments, teachers tend to look for criteria that are easy to apply--so that papers can be quickly graded with no head scratching--and in the case of large classes, even perfunctorily checked. Here is a typical handout of the criteria for weekly "think pieces." To be Acceptable the piece must be at least 750 words. You don't have to be right in everything you say about the course material, and you don't have to have a unified essay with a single point. But you must seriously wrestle with or engage the week's homework reading and the topic or issue that I specify. Informal, colloquial writing is fine, but it must be clear to me as reader. Handwriting is acceptable--even a few scratch-outs with new words or phrases written in--as long as the piece is neat enough to make it genuinely easy to read.

One could add other features to the criteria: for example, that students quote a passage from the reading and work with the quote; or that students be more or less right in what they say about the reading--or at least not badly wrong. (This criterion would require more careful reading by the teacher.) One could ask that pieces have a main point or be unified, but students often sacrifice genuine exploratory thinking when they are trying to make everything "fit a thesis."

In low stakes grading we can get away with some flexibility in criteria: we can slip the bar up and down a bit depending on the student. That is, we can easily insist that a skilled and well prepared student do a better job for Satisfactory than an unskilled student with a weak background who is trying hard and improving. When the stakes are low, students aren't likely to notice or to mind this kind of flexibility, as long as we make it possible for all students to get a Satisfactory without inordinate effort. (For I don't want to push students so hard on low stakes pieces that we lose the essential benefit: that we don't have to work too hard, and that students get to take chances, take risks, explore, do it their own way--yet do the work and engage the material.)

The criteria for high stakes assignments, however, can become a high stakes matter. That is, even though using fewer levels makes grading or sorting easier and more trustworthy--since the fewer the levels, the fewer the bars we have to set, the fewer the piles we have to sort papers into, and hence the fewer the occasions for uncertainty or argument or "wrong" decision-nevertheless, if the stakes are high there will be anxiety and potential argument around where we set those fewer bars or borderlines: between satisfactory and not satisfactory--or between satisfactory and excellent. So how do we decide how high to place the bar? How do we figure out exactly what we mean by "a satisfactory essay" or "an excellent essay"? What is good writing? Acceptable writing? This is where we get into head-scratching with ourselves, unpleasant arguments with students, and honest disagreement with colleagues.

Here's where we have to think hard about criteria. But this doesn't mean we have to make our lives impossible--especially if we are teaching a large class that doesn't center on writing. There is a traditional and crude distinction between Form and Content that actually works quite well (despite some criticism of it as old fashioned or even theoretically suspect). For example one can explain one's criteria for a high stakes essay as follows: "I will grade on a three level scale, Unsatisfactory, Satisfactory, Excellent, and I will count roughly two thirds for content and one third for form. By "content" I mean saying valid and interesting things about the course material."

But we use more detail if we want. We can split Content into, say: -Accuracy with the material we are studying and -Good thinking or reasoning. We can also split Form into: -Clarity and -Mechanics. Or we can specify criteria that are more particular to the assignment-intellectual operations that are central for different assignments. Perhaps one essay centers on analysis, another on persuasion--or research, or applying course concepts to new instances.

Spelling out criteria doesn't mean we have to explain in a comment on each paper exactly why it got the grade it got. It's when we don't spell out criteria that we have the most obligation to spell out individual reasons for individual verdicts. If our criteria are complicated in any way, we explain and describe them in a handout. (Nor does this obligation to think about grading criteria mean that history or biology faculty have to talk like writing teachers. Students do better if we use simple everyday language about writing, using words like "clear" or "tangled," rather than technical "English teacher" terminology.)

Simply to spell out our criteria in public--in class or on a handout--is far better than what happens with most grading, namely leaving the criteria tacit and mysterious. When we spell out criteria in public, we usually grade more fairly. That is, when we articulate and think about our criteria, we are not so likely to be unduly swayed if one particular criterion or feature in the writing is terribly weak or strong. (Research shows that teachers tend to get annoyed by papers that are full of mistakes in usage and full of nonstandard dialect, and consequently overlook virtues in content or reasoning in such papers.)

But the principle of minimal levels comes to our rescue here to show us that it is not in fact so difficult to do what is really most valuable: to give students individual feedback on their high stakes essays in terms of the criteria we are using. For just as it isn't hard to read through papers and merely pick out ones that are notably weak or notably strong, so it isn't hard merely to note if an essay is notably weak or strong on the criteria we have named as important. Thus the "grade" on the high stakes paper might look like this: COURSE CONTENT: excellent; THINKING: excellent; CLARITY: satisfactory; MECHANICS: satisfactory; OVERALL: excellent.

When we give this kind of feedback (and let me stress that we can grade adequately without doing so--especially in a huge class), notice that we are not bound to calculate the final overall grade on the paper by pure arithmetic. Most of us, for example, will decide that the content criteria count more than the form ones. This approach is clearer and fairer because it tells students the basis of our grade. By noting strengths and weaknesses on a crude scale on multiple criteria, we give far more meaning and clarity to three-level grades than students get from a ten-level conventional grades. Yet doing so is easier than trying to figure out that conventional grade--and it permits us if the class is very large to get away without writing a verbal comment. In fact, these crude notations on criteria are often more helpful than most verbal comments.

But perhaps even more valuable, students at last get some substantive feedback on what they did well or badly. Yes, most students will probably get Satisfactory, but these students benefit a great deal from knowing which dimensions of their paper were notably weak or strong. And the students with Unsatisfactory definitely benefit from more particular feedback--not just from having their most egregious sins named, but just as important, from getting some encouragement by seeing that not everything was unsatisfactory.

Notice a crucial principle here: criteria for grading really involve two dimensions. First, there is the vertical dimension: how high will we put the borderlines between pass and fail or between strong, satisfactory, and weak, or between the letters of conventional grades? And second, there is the horizontal dimension: what features are we going to count or evaluate (e.g., content, applying concepts to new instances, etc.)?

This distinction between the vertical and horizontal dimension of grading criteria may be the most productive thing to understand when it comes to high stakes grading. As with stakes and levels, these two dimensions are often tangled together, but they are different and can be manipulated separately. Most of the difficulties in grading occur in the vertical dimension. Everything works better for us and our students if we put less emphasis on it and instead emphasize the horizontal dimension. Conventional grading from A to F emphasizes the problematic vertical dimension: making fine-grained distinctions as to pure undifferentiated, unarticulated quality--and that's where all the disagreement and argument and unfairness come in. It's extremely difficult and problematic to describe levels of quality. (For an A, on this paper, "really good" is not good enough; it has to be "really really good"?) It's much easier and more educationally productive to describe the multiple features or dimensions that one is looking for in good writing (or any kind of learning) than to make simpler, cruder verdicts about quality.

In short when we really give a serious answer to the question of criteria--to "what we are looking for in quality"--we see that there is a clear choice in the evaluation process. We can emphasize the vertical dimension. This means making a single difficult evaluative verdict along a single undifferentiated scale that has multiple fine gradations. Or we can emphasize the horizontal. This means figuring out what one is looking for and then making multiple simpler, cruder evaluative verdicts on multiple criteria.

The Killer Criterion: Mechanics or Grammar and Spelling

In trying to deal with this disputed criterion, I find it helpful to distinguish between correctness and clarity. My position is that we can and should insist that students achieve both, but in different ways. That is, if students are writing in a high stakes situation where they cannot take time or get help (such as on an exam), it seems to me nevertheless fair to demand that they be able to explain the concepts or operations of history or biology clearly and readably in writing, but not to demand that they be able to avoid grammatical and spelling mistakes in this writing. An exam essay can be clear and readable and still have awkward or ungainly or unidiomatic sentences and quite a few mistakes in grammar and spelling. If the essays or answers on an exam like that are excellent, I'd grade it Excellent. But if the language is so tangled as to be genuinely unclear, then that seems to me to hold the grade down. (And for low stakes pieces they write at home, I also insist on a certain minimal level of clarity but not correctness.)

But for high stakes assignments that they do at home--where they can get help--I believe it makes sense to ask not just for clarity but for correctness too. Even from students whose first language is something other than English or a dialect of English. The truth is that most writers get help in revising and copy editing when a piece of writing matters. Some get it from friends; some pay for it. I want to push my students to get this help. It's most obvious that ESL and dialect speakers need to learn how to get the help they need so as not to hand in writing that will put them at a disadvantage. But in fact these students tend to be aware of their need for help in copy editing. It's the native speakers of standard English who need to have the tough criterion strictly enforced--and often feel that it is unfair to do so.

(But let's not be fussy in our standards for correctness. Notice that the sentence italicized above is strictly speaking a "fragment": it lacks a verb. But it's the kind of fragment that is pretty well accepted in published writing now.)

I'm making two distinctions here. First, I'm distinguishing between clarity and correctness. Clear writing, clear explanations: these are plenty hard to achieve. We will be more successful in demanding them if we don't get them mixed up with correctness.

Second, I'm distinguishing between achieving correctness with and without help. I don't think we can demand correctness without help, but surely we do want to insist vehemently that all students be able to get their writing correct with whatever kind of help they need. They have to do whatever is necessary to turn in clean copy, and they won't do it unless we make it a condition for credit.

But these are arguable matters and people take different approaches. For example, some teachers who take a more lenient stand say that you can get a satisfactory paper if there are lots of mistakes but not an excellent one.

Conclusion

There are two important principles I want to emphasize: (1) That there is a crucial difference between how high the stakes are on an assignment and how many levels one uses in grading. (2) That in spelling out criteria, we have a choice between emphasizing vertical differentiation of quality and horizontal differentiation between features of writing that we are looking for. I think that awareness of these principles can help a teacher work more clearly and coherently toward any grading policy. But let me summarize my own position more nakedly.

I think teaching and learning suffer from too much emphasis on numerical grading. At times I would like to do away with it altogether. I taught for nine years at Evergreen State College where the learning and teaching climate was vastly better than usual because there were nothing but written evaluations- and still students did well in getting into professional and graduate schools and getting jobs. (People don't much trust standard grades.)

But we can make big improvements by making much smaller, less drastic changes:

My premise is that we don't need all those ten levels of distinction between F and A--especially since no one knows what conventional grades actually mean and they are not trustworthy. But we do probably need to be able to send the main messages and send them explicitly in words: "this is definitely acceptable or satisfactory college level work"; "this is unsatisfactory work"; "this is excellent or outstanding work." And unless we have too many students, we need to point to features or dimensions in someone's work that are notably strong or weak.

But we need to recognize that these "strongs" and "weaks" don't translate into fine-lined gradations such as conventional grades give us between B-plus and B-minus. Which is a better essay or better writing? Something clear and well organized but that just summarizes the ideas in the textbook or something with lots of interesting thinking that goes beyond the textbook but that thereby makes some genuine blunders? My point is that there is no answer: it depends on what qualities you care most about. Under conventional grading, some teachers give a higher grade to one essay, some to the other; but no one knows anything about what these teachers actually mean or anything about the actual strengths and weaknesses of the students.

Arguments about grading tend to bog down into fruitless either/or debates: yes or no, pro or con. I hope my analysis shows that we can get out of this binary realm. There are various kinds of "minimal" or "almost" or "sort of" grading--and a wide range of ways of deploying the elements that make up grading.

------------------------------
Peter Elbow is Professor of English Emeritus at UMass Amherst and former Director of the University Writing Program (1996-2000).

This essay appears in Writing to Learn: Assigning and Responding to Writing Across the Disciplines, 1997, a volume of the Jossey-Bass Teaching and Learning Series. Edited by Peter Elbow and Mary Deane Sorcinelli.

Updated September 3, 2008

 

UMail / UDrive / Spark / Spire