Scaffolding a research paper across a semester

The term paper has a failure mode everyone knows: assigned in week two, mentioned in week eight, written in the seventy-two hours before it is due. The instructor reads thirty papers that are first drafts, and the students learn that a research paper is a thing you produce in a weekend, which is the opposite of what the assignment was for.

Scaffolding is the fix, and everyone knows that too. What is hard is finding deliverables that are real — that a student cannot fake in an hour the night before, that the instructor can read in five minutes, and that add up to the paper rather than to a pile of paperwork beside it. Cards give you those deliverables for free, because the card method (student guide 1) already has stages, and each stage leaves a stack in a particular state.

This guide is a fourteen-week scaffold built on those states. Six deliverables, one conference, one paper. Every deliverable is something the student already has if they are doing the work, and nothing if they are not.

The principle: hand in the stack, not a report about the stack

The traditional scaffold asks for documents about the paper: a proposal, an annotated bibliography, an outline. Each is a piece of writing in its own right, each can be produced from nothing the night before, and none becomes part of the paper. Students learn to write proposals.

The card scaffold asks for the stack itself, at a stage. A student who has been reading has source cards; one who has not, has none. There is nothing to write up, nothing to fake, and every card handed in at week four is a card that goes into the paper at week fourteen. The deliverables are not beside the work. They are the work, photographed in passing.

The scaffold

Total reading beyond the paper itself: about twenty minutes per student across the term, most of it in the conference, which is the part worth the most.

Each deliverable, and what it teaches

Week 4, source cards. The first deliverable is the smallest and it does the most: it moves the reading to week three. Eight source cards with full citations cannot be produced without having found eight sources, and finding them is the part of research that students most postpone. Your five minutes per student here is spent on the set: is it one book and seven websites? Is it eight things from the first page of one search? Say so on a card and hand it back. This is the cheapest intervention in the whole term, and it changes the paper more than any comment on a draft.

Week 6, evidence cards. Quotations exact, with pages; facts in the student's words, with pages. What you are checking is the discipline from student guide 3 — one thing per card — and the source numbers, because a card without one is a citation error you will otherwise meet in week fourteen. Expect thirty to sixty cards. You do not read them all; you fan them and look at ten.

Week 8, the row. This is the deliverable that shows whether an argument exists. The student lays out claims with evidence under each and counterpoints beside, and photographs it. You look at the photograph and ask the three questions the layout answers (instructor guide 1, minute 22). Most rows at week eight have a claim with nothing under it and an objection with no answer, and that is fine — it is week eight — but now both are visible, to you and to the student, six weeks before the paper is due instead of six hours after.

The conference. Ten minutes, the cards on the table between you. This is the whole scaffold's reason for existing, and it costs five hours for a class of thirty. Do not look at prose; there is none. Look at the row and ask: which claim are you least sure of? What would the smartest person who disagrees with you say? What do these claims have in common — can you say it in a sentence? The student leaves with two or three cards to write and one to cut. A student with no row at the conference has a problem you can both see with six weeks to fix it, which is the other reason the conference is here and not at week twelve.

Week 10, the thesis and two orders. The thesis card at the top of the row (student guide 4), and two photographs: the row in two different orders, with one sentence saying which is better and why (student guide 6). The sentence is what you read. The counterpoint first, because otherwise the reader is arguing with me for three paragraphs is a student who has understood that order is a choice. The second one, it flows better is a student who has not yet, and the note to hand back is why?

Week 12, the draft. Written from the chosen order, one claim per section, bridges between. You read it as you read any draft, but you have the row beside it, and the comment your draft's second section argues a claim that is not on your row is a comment you could not have made before.

Week 14, the paper and its record. The paper, and the stack: the dated cards, the cut pile with its reasons, the photographs of the orders. You read the record the way instructor guide 2 says — for shape, not numbers; assuming the pattern is a person; to ask, not conclude — and for most students you will not need to read it at all, because you watched the paper get built.

What to grade

The paper. Each deliverable is complete or not, on time or not, and that is all; a rubric for source cards produces students who write source cards for the rubric. The scaffold's effect shows up in the paper, and it shows up as papers that were written across twelve weeks by people who knew what they were arguing by week ten. That is what you grade, and it is better than what you were grading before.

If a student misses a deliverable, the conversation is where are your cards, and the answer tells you what you need to know in a way that a late proposal never did.

Variations

A shorter term. Ten weeks: sources at 3, evidence at 4, the row at 6 with the conference in the same week, thesis and orders at 7, draft at 8, paper at 10. The conference is the piece to protect.

A larger class. The conference is the cost that scales. Two options that keep it: pairs — two students, twenty minutes, each reviewing the other's row while you listen (instructor guide 3 is the workflow); or a teaching assistant running week eight's conferences from the three questions, which they can do, because the questions are about a table and not about prose.

A discipline with its own conventions. The stages are the same in history, sociology, biology and literature; what differs is what counts as a source card and an evidence card. Say what counts in week two. A lab science's evidence card may be a result rather than a quotation; the row is the same row.

If the class uses software

Software that works from cards makes every deliverable a thing the student already has and can hand in without photographing a table: the sources with their citations filled from identifiers, the evidence cards linked to their sources, the row as an order the software keeps, two orders kept side by side and compared, the draft compiled from the chosen one, and at the end a record of the process generated from the work itself, dated as it happened. It also asks the three week-eight questions on demand — a claim with nothing under it, an objection nobody answered, a source never cited — so the student can ask them before the conference instead of hearing them at it.

What it does not do, and should not: show you any of this before the student hands it in. There is no view of who has started, no activity feed, no count of cards per student. The scaffold works because the deliverables are handed in, at the stages, by the student; a scaffold the instructor could watch from a dashboard would be a different thing, and a worse one, and the software built on these principles does not build it. The ten-minute assignment setup and the syllabus handout say this in the words to give students.