Information Sheet #2

September 15, 1987


In the NCTE publication Research in Written Publication (1986), George Hillocks, from the University of Chicago, reviews recent studies that examined the effect of teacher commentary on student writing. Hillocks concludes "that the results of all these studies strongly suggest that teacher comment has little impact on student writing."

Hillock notes, however, that some studies tentatively suggest how the profession's failure may come from the diffuse, unfocused commentary students receive from teachers. Commenting procedures which concentrate on one or two key aspects of writing appear to have a much better chance of helping writers improve. One such system that two or three teachers at Coe use is called "Minimal Marking." The following description of such an approach is from an article by Richard Hasswell of Washington State University. The article originally appeared in College English (Oct, '83).

It is a disturbing fact of the profession that many teachers still look toward the marking of a set of compositions with distaste and discouragement. Reasons are obvious, not the least being the intuition that hours much be put in with little return in terms of effect on the students or their writing. C.H. Knoblauch and Lil Brannon's survey of the research on the effect of marking unfortunately supports this intuition. Positive results of teacher intervention through written commentary simply have not yet been found. The problem is analogous to that of the teaching of grammar in composition courseshundreds of thousands of hours spent on the task of little proven benefit. Fortunately, however, Knoblauch and Brannon balance their description of unfruitful paths with a model of paths still promising.

Whether Knoblauch and Brannon's model of beneficial written commentary can be verified by research remains to be seen, but I would like to provide evidence here that suggests it will be. In essence they propose commentary that 1) facilitates rather than judges, 2) emphasizes performance rather than finished product, 3) provides double feedback, before and after revision, and 4) helps bridge successive drafts by requiring immediate revision. All these requirements are met by a method of marking surface errors in writing that I have been using for several years and recommending for use by teaching assistants.

My own application is as follows. All surface mistakes in a student's paper are left totally unmarked within the text. These are unquestionable errors in spelling, punctuation, capitalization, and grammar (including pronoun antecedence.) Each of these mistakes is indicated only with a check in the margin by the line in which it occurs. A line with two checks by it, for instance, means the presence of two errors, no more, within the boundary of that line. The sum of checks is recorded at the end of the paper and in the gradebook. Papers, with checks and other commentary, are then returned fifteen minutes before the end of class. Students have time to search for, circle, and correct the errors. As papers are returned to me I review the corrections, mending those errors left undiscovered, miscorrected, or newly generated. Where I feel it is useful, mistakes are explained or handbooks cited. Within those fifteen minutes I can return about one third of the papers in a class of twenty-five, and the rest I return the next session. Until a student attempts to correct checked errors, the grade on the essay remains unrecorded.

The simplicity of this method belies its benefit. First, it shortens, gladdens, and improves the act of marking papers. Because the teacher responds to a surface mistake only with a check in the margin, attention can be maintained on more substantial problems. The method perhaps goes a long way toward dimming the effect of surface mistakes on evaluation, since much of this negative influence may arise from the irritation that comes from correcting and explaining common errors (its and it's) over and over. On the second reading the teacher does not lose the time gained initially, for according to my count students will correct on their own sixty to seventy percent of their errors. The method saves me about four minutes a paper. That is nearly two hours saved with a set of 25 essays.

Second, the method forces students to act in a number of ways that have current pedagogic sanction. In reducing the teacher comment on the page, it helps avoid the mental dazzle of information overload. It shows the student that the teacher initially assumed that carelessness and not stupidity was the source of error. It forces the student, not the teacher, to answer the question, and reinforces learning with successful solutions. It engages students in an activity that comes nearer to the very activity they need to learn, namely editingnot the abstract understanding of a mistake someone else has discovered, but the detection and correction of errors on one's own. Finally, improvement is self-motivated. The fewer mistakes students submit, the sooner they leave other students still struggling in the classroom with checks by every third line.

Category of Error

Number of errors checked in margin by teacher

Number or errors correctly emended by Students

Percent corrected by students

Semantic signaling (capitals, underlining, " " and ' marks)




Syntactic punctuation




Spelling (including hyphenation)




Grammar (including tense change, word omission, pronoun disagreement)




All errors




[Reproduced above is] the breakdown of correction that 24 freshman in one of my recent classes made on their first in-class essay (without recourse to a dictionary). Crude as this breakdown is, a useful fact immediately emerges. Students are able to find and correct different kinds of errors at about the same rate. In short, more than half of the surface errors students make, regardless of type, occupy a kind of halfway house between purely conceptual and purely performance-based (only a few seem truly slips of the pen). They are threshold errors, standing on the edge of competence in an unstable posture of disjunction ("I know it is either conceive or concieve") or of half-discarded fossilization ("I don't know why I capitalized 'Fraternities.' I know that's wrong."). It is good for the teacher to be reminded that, after all, the majority of errorsall kinds of errors, and differently for different students"mark stages on route to mastery." Further, the method isolates, for each individual student, those errors of deeper etiology. It is remarkable how often the method winnows away a heterogeneous clutter of threshold errors to leave just a few conceptual errorserrors, though again idiosyncratic and multiplied by repetition, not accessible for focused treatment. So the method is an ideal first step in the pedagogical attack on errors recommended by Paul Diederich and others: keep records, isolate a few serious errors, individualize instruction....

The ultimate value of this method for me is that it regulates what I consider a minor aspect of the course to a minor role in time spent on marking and in class, while at least maintaining and probably increasing the rate of improvement in that aspect. Crudely put, less work for the teacher, more gain for the student. But the gain may be compounded in ways more complex than this suggests. Knoblauch and Brannon rightly point out that commenting must be evaluated in terms of the "full teacher-student dialogue." Now too much commenting can harm this dialogue in at least two ways. It will embitter the teacher with the knowledge that the time and energy spent on it is incommensurate with the subject and the results. And it will frustrate both teacher and student because judgmental commentary unbalances the teacher-student equilibrium in an authentic learning situation, that is, where the student is doing most of the work. Long ago Comenius put it best: the more the teacher teaches, the less the student learns. This marking technique postpones correcting, emoting, and describingwhere the teacher does all of the workand instead suggests, questions, reminds, and assigns. Because students do most of the work, the discouragement of which I first spoke subsides, and a certain freshness and candor return to the dialogue.

Can this method be transferred to other aspects of the writing? I think so, although right now I must speculate. Certainly problems of writing that lend themselves to spot improvement could well be marked with marginal checks: injudicious diction, needed transitions, unsupported generalities. Larger, structural problems such as stumbling introductions and disordered paragraphs might be signaled with marginal lines. More interestingly, so might fallacies and other lapses in thinking. In each case the effort would be to find the minimal functional mark. The best mark is that which allows students to correct the most on their own with the least help. An obvious pedagogical truthbut one that runs counter to the still established tradition of full correction.

This website created and maintained by the Coe Writing Center. Copyright 2001.
Email Dr. Bob Marrs with any questions, comments or suggestions.