You already read this essay once. You left comments, the student revised, and now the revision is back on your desk asking for another full read at the same pace as the first one. If you're rereading every sentence cover to cover a second time, you're paying full grading price for work you already partly graded.
Revision grading should cost you time in proportion to what actually changed, not the length of the whole paper.
Grade the delta, not the document
The fastest way to do this is to have the student make your job easier before it reaches you: require a short revision note attached to the resubmission, naming exactly what they changed and why.
- "I moved my thesis to the end of paragraph 1 because you said it was buried."
- "I added a counterargument in paragraph 4 — didn't have one before."
- "I cut my last paragraph and rewrote the conclusion instead."
With that note in hand, you read for two things: did the named change actually happen, and did it actually fix the problem you flagged. That's a targeted read, not a cover-to-cover one.
Using AI to spot the actual diff
For a longer essay, tracking exactly what changed between two drafts by eye is slow even when you know roughly what to look for. This is a place AI is genuinely fast and low-risk, because you're using it to locate differences, not to judge the writing.
Try this:
- "Here's my student's original draft and their revision: [paste both]. List the specific paragraphs or sentences that changed, added, or were cut. Don't evaluate quality, just show me the diff."
- "My original feedback was: [paste your comments]. Based on the diff above, which of my specific comments did the student actually address, and which ones weren't touched?"
That gives you a map before you read a single sentence yourself: here's what moved, here's what's still open. You read those spots closely and skim the untouched parts to confirm they're still solid, instead of treating the whole paper as new.
Keep your second-pass rubric shorter than your first
Your original rubric graded the whole essay. Your revision pass doesn't need to re-score everything — it needs to score whether the specific feedback got addressed, plus a quick check that the rest of the paper didn't fall apart in the process. A two-column revision rubric (issue named in round one → resolved / partially resolved / not resolved) is faster to fill out and more honest about what you're actually re-evaluating.
Our [rubric generator](/grading-rubric) can build that lighter revision-specific version straight from your original feedback comments, so you're not writing a whole new rubric from scratch for the second pass. And when it's time to hand feedback back again, our [feedback tool](/student-feedback) helps you write comments that point at what's still unresolved instead of repeating the same notes from round one.
What still needs your judgment
- Whether a "revision" is actually a revision. A student who reworded three sentences and called it a full rewrite needs a conversation, not a passing grade on the resubmission.
- How much credit partial fixes earn. A thesis that moved but is still vague isn't a zero and isn't a full fix — decide your partial-credit line before you start grading, not paper by paper.
- When a second revision cycle is worth it. Not every essay needs a third pass. Save that offer for students where one more round will clearly move the grade, not as a blanket policy.
