In my philosophy courses, I largely give take-home writing assignments. These are either focused on the course readings (no research required beyond the syllabus) or not (so they require students to do outside reading). Usually, I have 5-7 of the former (1,000-1,500 words) and 1-or-2 research papers (2,000-2,500 words if it is one long paper, or 1,000-1,500 if two shorter papers).

These assignments always have as their target audience other students who have not read the same material, not me, and the tasks are typically as follows:

1. Clearly state the argument in the text with reference to the text (e.g. quotes or paraphrase, in either case with citations to the reading and (maybe) supporting materials like instructor notes).
2. Clarify key concepts or critical applications of some part of the argument using both a reader or instructor-provided definition and some student-selected example or imagined scenario.
3. Raise and (maybe) address one or two critical concerns about some part of the argument.

My goal in assigning students these tasks is to assess whether they have understood some part of the text, typically taking this to mean that they can critically explain what they have read to a peer by doing 1-3 above.

I use a two-draft model for most take-home assignments that I want to explain and motivate in this post. Whether a short writing assignment or a research paper, I make students do the following:

1. Submit a complete (or at least substantive) first draft for a grade
2. Submit a cover page with their final revised draft that…
   • Summarize, in the student’s words, the main suggestions from my feedback on their first draft
   • Summarize in the student’s words what revisions were made in response to my feedback

This forces students to:

• Read my feedback, which is something that is not done by a vast number of students for a one-and-done assignment (there is little if any grade-related reward to read instructor feedback on an old assignment in such cases)
• Critically think about my feedback (which often leads them to come to talk to me during student hours to ensure they are responding appropriately)
• Critically think about what they said (which often leads them to write something better, say, with better examples, changing their thesis to something they knew or actually believed, and so on)

This helps me:

• Assess their learning more conversationally
• Signal that their assignments are works-in-progress and that perfection is not the standard
• Signal that their assignments are participating in a conversation (whether well or poorly) because they are forced to converse with me via my feedback (if not also in-person)
• Give them a second chance if they flubbed the first draft (“two bits at the apple” and all that) since students are learning—if they all knew what I was teaching, what would be my use?

Let me now address some concerns you may have about the two-draft assignment model.

How do you ensure students are not cheating if no part of the assessment is done with an in-class assignment like an exam?

You can do a two-draft assignment model with exams, too, if you have students revise their old answers and submit the revised versions. The two-draft model isn’t wedded to take-home assessment, though I tend to prefer take-home assessments so that we can have more discussion in class.

Aren’t you worried about genAI programs like Chat-GPT4 and using those to write papers?

No. The two-draft model is hard to passably follow with plagiarism generators (a.k.a. “genAI”). Sure, widely available plagiarism generators can produce a poor first draft that uses made-up page numbers and baseless summaries of the text. (For that sort of concern, just make your students cite the text in their answers and also cite your course’s version). But such programs have a hard time responding to comments on a PDF in any coherent or sensible way, much less writing a nice cover page in response to it, much less making tracked revisions to a document.

Isn’t this a lot of work?

Yes—up front. But it actually saves you tons of time when you grade the final draft. You are assessing something you have mostly read already and you have your first-draft comments to guide evaluation of the students’ cover page and final draft. My exam weeks are largely stress-free (and yours can be, too!). I don’t say that to brag, but to highlight what might be possible in your life as well as in mine.

Nice of you to say, but you are tenure-track and I am not. Isn’t this too much work for me even if it may work for the privileged few?

Yes. Those teaching the equivalent of five-plus courses (or 150-plus students) each semester are overworked plain and simple (and almost surely not compensated justly for their quite valuable labor). Probably the “overworked” number(s) can be set even lower than that, but where we should set that bar is not the point here—as a matter of routine practice, it is way too high, to the misfortune of students and professors. The point is that many faculty in philosophy (and not just in philosophy) are asked to effectively teach far more people than is reasonably possible in the fixed amount of time that we have on a given day. This makes giving the detailed feedback that is required for the two-draft model to work practically impossible for some faculty. This is not the fault of the two-draft model: a two-draft model (indeed, any assignment model) is not designed to fix exploitative labor practices, nor to fit with them. Such conditions are to be survived or, ideally, dismantled—not accommodated (including by acting like our best pedagogical practices really fit the exploitative behaviors dominant in academic workplaces).

Isn’t this just scaffolding, or a less fully sequenced (and perhaps worse) version of it?

No. This does involve (and pair naturally with) assignment scaffolding, given that there are a minimum of two graded or assessed drafts. But the cover page is I think (I might be wrong) less commonly used. Scaffolding alone doesn’t imply anything about cover pages or grading a substantial portion of a paper twice (or more). Writing assignments can, consistent with the two-draft model, be scaffolded in various ways. I have run various sequences of the model, some with in-class work (presenting final papers or projects, brainstorming papers with pair-and-share, etc.) and some with out-of-class work (submitting a topic idea or outline for a proposed paper, say). But the core of the assignment model is the two drafts with a cover page, whatever sequencing happens before, during, and after a first draft is assessed. The main thing that I am promoting here is not scaffolding in general (which many like myself think is an effective structure for assessments). What I am focusing on here is the second draft with a cover page. The exercise of making students explain my feedback to me, and summarizing how they responded to my feedback retrospectively after completing the final paper, has improved their final drafts and made (in my experience) for much easier grading of final papers.

Can’t students explain your feedback and how they responded to it in the final paper in other forums, like during an “oral exam” or similar (i.e. a conversation that is somehow graded)?

Maybe they could. I’m not in principle opposed to that, though I haven’t tried it. I like telling students to sit down and write up what they thought I suggested and what changes they plan to make (or not) in response to those suggestions. But I don’t see that there is anything sacred about writing it all down, so long as the student is forced to do 1-3 above. A conversation before submitting a second draft, and/or after submitting the final draft, so far as I know, could achieve those same goals. Give it a go.

Does this model apply to teaching logic or quantitative reasoning courses with a PHIL prefix?

It doesn’t obviously do so—those courses generally involve learning different skills, much more akin to training the mind’s general faculties (by giving the student instructor-determined tools—like proof rules—and asking them to complete specific tasks—like evaluating an argument’s validity). That is easier to assess on a one-and-done exam because learning to do logic is like learning to lift weights—it takes repetition and practice. However, it strikes me as readily possible to adopt a two-draft model in such classes—I have not yet attempted this but I might in the upcoming semester—by asking students to revise their earlier answers and attempts with comments on what they previously did wrong.

Did you address all my concerns?

No, I probably did not. But any and all comments, concerns, criticisms, questions, and suggestions about the two-draft model are most welcome. You can contact me here.

The post The two-draft assignment model first appeared on Blog of the APA.

Read the full article which is published on APA Online (external link)