Wednesday, March 17, 2010

The merits of collective feedback

This quarter, I tried a new approach to giving students in my General Education courses feedback on their assigned essays. I think it went OK, but it met with grumbling among some students. So I'd be interested to know how I might make the approach more effective and more satisfying to students.

The background: This quarter, I had 123 students in my intro to ethics sections. This is about 50% more than I'm accustomed to having. I assign them two 1,000-word essay assignments during the quarter. (They are required to revise one of these essays and turn it in at the end of the quarter.) Needless to say, the grading burden for me is great. Just reading 120,000+ words is time consuming enough. Evaluating the work and giving feedback only adds to the workload.

So here was the approach I took: I read about one-quarter of the student papers carefully, and took a quick glance at the other three-quarters. (Since the students can choose to write on different topics, I made sure that the sample of papers I read carefully reflected the number of papers written in response to each topic.) All the papers were evaluated on a pass/fail basis: A paper passed if it was broadly on-topic and suggested minimal effort. Each of the two papers was 5% of the students' quarter grade (the revised essay was 25% of the quarter grade). I then developed collective feedback based on the one-quarter of the papers I read carefully, highlighting common organizational, logical, and interpretive problems in these papers. I provided both general feedback, based on all of the papers I read carefully, as well as feedback specific to each topic. I then made all of this feedback available to students via Blackboard.

So here are the merits to this approach as I see it.
  1. It obviously reduces my workload, as well as decreasing the turnaround time for students. If I had graded and given feedback on each of the 123 papers, it would have taken me at least ten days, I suspect. With this approach, students get a pass/fail grade and some feedback within five days.
  2. I use rubrics and comment codes on papers, but I still find I make many of the same comments on student papers anyway, so this approach saves me the time of repeating the same point, say, fifty times.
  3. The pass/fail grading essentially turns the essay assignments into first drafts. As a result, they are relatively low risk, but they prepare students for the more high-risk task of revising and resubmitting one of the assignments later on. It also has the positive benefit that students' grades end up turning more on their best performance rather than on their worst, i.e., the grade reflects their capabilities. In an introductory philosophy course, where students are confronting something wholly new to them, letting their best work represent their learning (instead of their early struggles) strikes me as equitable.
  4. This approach also requires them to revise, which I believe to be at least as important to the overall writing process as initial drafting.
  5. It sends a realistic and accurate message to students about the economic situation within the public university system I teach in. Faculty workloads are increasing, and our educational model is increasingly off-the-rack rather than customized. Wholesale feedback for wholesale education!
  6. It makes the students do some self-diagnosis with their writing in that they have to identify which elements of the collective feedback apply to their own essays and which don't. My hope is that they then have to look at their own writing with a careful and discerning eye. Ultimately, this self-diagnosis is part of being a skilled and autonomoous writer, so my aim is to give them practice at such self-diagnosis. Indeed, one thing I worry about with individualized feedback is that it is too helpful in a sense. I the instructor diagnose the problems in the paper, the student 'fixes' the problems, and the process is over without the student really developing the capacity to see their own writing from a critical perspective. This 'jump!'/'how high?' dynamic runs against an important long-term goal of teaching writing.
As I said, I think this approach worked OK, but there was grumbling from a few. The grumblers wanted individualized feedback. Some weren't sure about how to apply the collective feedback to their own work (point 6 above). But since I think this approach is pedagogically sound, as well as being in my own considered interests, I'd be interested to know how the approach could be tweaked or modified.


  1. I'm definitely intrigued by this clever solution, Michael. Could you please give some examples of the sorts of comments you posted on Blackboard? It would help me assess the proposal a little better, and I'm interested in assessing it for my own use. Thanks!

  2. Having been on the receiving end of this kind of feedback before, I would wager that there are far more disgruntled students than you realize. If you didn't collect feedback from everyone on this specific aspect of the course, I'd encourage you to put together a quick SurveyMonkey survey and email your students about it.

    That said, I certainly understand your reasons for taking this approach. I think the pass/fail grading is an essential element, both to save time and mitigate student frustration.

    Four suggestions:

    (1) I assume you do this, but if not, encourage students to come meet with you for individualized feedback.

    (2) Post some samples of good papers for each topic (with permission, of course).

    (3) I couldn't tell from your post whether you're giving them "completed" rubrics. If not, I'll bet you could do so with minimal extra effort. I use a "grid rubric." I put criteria (e.g., thesis statement, argumentation) along one axis and ratings (e.g., excellent, unsatisfactory) on the other. I fill in the cells with a description of, e.g., an unsatisfactory thesis statement. I highlight the appropriate cell for each paper and give students the completed rubric. (I also give individualized feedback, but I've never had a class anywhere near 120 students.) If you know enough about a paper to give a pass/fail grade, you probably know enough to fill out a rubric like this.

    (4) Give students an (optional? extra credit?) assignment in which they have to explain in detail (a) how the collective feedback applies to their paper, and (b) how they would revise their paper in light of that feedback. This, too, could be graded pass/fail. I've tried this before with short assignments, and it worked quite well.

  3. Anon 2:31: So here's a sample:

    First, some general feedback for all papers:

    "For the most part, these essays were better than those written for assignment #1. Some recommendations:

    * Quick, to-the-point introductions are best. Don't spend time telling us about how ethical theorists disagree, etc.
    * Read the relevant material - know of what you speak. Many of you plainly didn't understand the theories and /or objections you were writing about. Don't just guess what an objection amounts to.
    * Use transitions and signposting to guide your reader through your paper: "having explained the absolutist theory, let us now consider the objection that absolutists have no way to resolve conflicts among basic ethical rules," etc. Tell us where you're going so that we can follow along!
    * Think critically and sympathetically. Far too many papers had the following structure: here's a theory, here's an objection, theory refuted. That's too cheap and easy. Put yourself in the shoes of the theorist who's trying to defend a theory. How might they defend that theory? What strategies might they take?"

    Some feedback on a specific paper topic (this concerned the objection to contract theory that it cannot make sense of duties to non-human animals):
    "Contract theory and duties to animals:

    Papers written in response to this prompt were generally cogent and well-organized. Some specific observations:

    ·It's important to be clear about why there don't appear to be obligations to animals under contract theory. Note that the complaint can't be that animals can't actually agree to ethical rules. For as we noted, contract theorists generally suppose that what makes ethical rules binding is not our actually agreeing to them, but that we would hypothetically agree to them. If what the theory required was actual agreement, then we wouldn't even have ethical obligations to other people! The problem is rather that animals cannot even hypothetically consent to ethical rules and so are left out of the contract.

    • A slightly different way to explain why animals are left out of the contract: Remember that contract theory understands ethics in terms of reciprocity. Ethical rules are meant to serve each person's self-interest if everyone obeys those rules. In a sense, contract theory is a kind of enlightened version of ethical egoism. By constraining our self-interest according to certain ethical rules, the contract theorist argues, things turn out better for us from the standpoint of self-interest. But wouldn't we humans have no incentive to agree to ethical rules protecting animals from mistreatment? After all, they can't be expected not to mistreat us or to obey ethical rules. It simply doesn't appear to benefit us to have rules requiring animals be treated kindly. (This approach might also help illuminate some problems for the 'proxies' response to this objection.)

    · Some of you tried to argue that animals have the same emotions, feelings, attitudes, etc. as human beings. You don't need such a strong claim to make sense of ethical obligations to animals though. They don't need to be exactly like us in order for us to have obligations to them - they just need to resemble us in some ethically relevant way."


    Thanks for your thoughts. I'd be interested to know why I may be underestimating the disgruntlement. I do follow your suggestions (1) and (2), though I'm thinking that it might be better with (2) to post some papers that are high quality/medium/low, since that might actually facilitate the self-diagnostic skills better than posting the best papers. (4) is an excellent idea -- on my quarter system, time would be an issue there though.

  4. One thing I dislike about this strategy is that it gives your best students nothing but a few college credits. I wonder if it would be considered unfair to add a third tier to your grading system: pass-with-skill-or-effort, pass-with-low-effort (or "get by"), and fail. Then perhaps those in the top tier would "earn" more substantive feedback. If this would give more students initiative to do more than get by, it might be better for everyone.

  5. Michael,

    I'm not sure how to interpret your first question. Do you want to know what my evidence is for thinking that there's more discontent than you see? Or are you wondering why people might be disgruntled?

    I'd almost suggested posting high, medium, and low quality examples. I think it would be discouraging to students to have their work posted as an example of a low quality paper. If you had time, maybe you could write some "composite" low-quality papers that combine the most common flaws? I think you could get away with a "medium quality" if you also posted some specific advice about how to improve the paper; you'd be buying the student's approval with individualized feedback.

  6. David, I suppose I was asking what experience you have that suggests students may be more disgruntled than they let on.

    Kevin, I don't understand your comment. It suggests that I'm giving students pass/fail for the class, which I don't (and can't) do.

  7. Sorry, I didn't notice in my first reading of your post that the quick-glance papers count for only 10% of the overall course grade. I withdraw my objection.

    But now I wonder: What accounts for the remaining 65% of the grade?

  8. Point 6 is the one that tips the scale in favour of this style of marking. I've given detailed feedback only to have only the corrections I've noted corrected in the final draft - they don't generalize the feedback to the entire paper. I've also had students misunderstand a theory, prompting me to write pages of clear explanation before their seminar, which gets entirely ignored when they present.

    My concern is that many won't entirely grasp how the general comments apply to them specifically. If they pass, they might ignore any attempt at improvement even if they just barely pass. How would they know? Do you ever have students hand in the same paper convinced that since it passed the first time, it will be fine as a final draft?

  9. Why not assign one, rather than two, papers? In a class of 120, students can't really expect detailed comments on two papers.

    As a student, I'd be pretty upset if my paper weren't read more carefully than you describe for the 3/4s group. But, in a gen ed course, I think I'd be pretty happy to have comments on one paper, especially if I could rewrite it.

    I doubt that the comments from one large class will be substantially different from term to term. Why not hand out a sample sheet of guidelines in advance? If you already do this, and the students still make errors, isn't that a reductio on your post-paper comment strategy? (I mean, if they aren't heeding advice in advance when their grade depends on it, why think that they will heed it afterwards, when only their learning depends on it?)

    If you want to grade lots of students without much effort, make multiple choice exams, or short answers. They're a pain to write, but with minimal tweaks, you can use them repeatedly. A good one can be a fine evaluative tool.

    I just think it's preferable to do less better than more worse.


If you wish to use your name and don't have a blogger profile, please mark Name/URL in the list below. You can of course opt for Anonymous, but please keep in mind that multiple anonymous comments on a post are difficult to follow. Thanks!