Fifteen Minutes of 'I Liked It'
Mr. Delgado's ninth graders write a timed persuasive essay every two weeks, and district rubric scores have not moved across three cycles. Reading their drafts, he finds that nearly all revisions are surface-level: correcting spelling, adding a sentence at the end, swapping 'good' for 'great.' Peer-response sessions take 15 minutes and yield comments like 'I liked it' and 'add more detail.' He teaches 32 students in 45-minute periods, and the department requires four major essays a year; he has just finished the second.
1. What is the most likely explanation for the flat scores, and what evidence would you gather before committing to it?
Read the worked answer
Most likely explanation: students hold a definition of revision as correction. If revision means 'fix errors,' then the only available moves are surface moves, and their behavior is rational given that definition. The 'I liked it' peer comments are consistent: no one has given them criteria, a procedure, or a reason to look at ideas and organization.
Before acting, check three things:
- His own feedback. If his marginal comments are also mostly mechanics, students are accurately reading what the system rewards. Sort his last set of comments into surface and meaning-level.
- The rubric. If the rubric is used only for a final grade and never mid-draft, students have no target.
- Whether students can even see a difference. Give them two anonymous paragraphs-one with weak evidence, one strong-and ask which is better and why. If they cannot articulate it, the gap is conceptual, not motivational.
What would change the diagnosis: if students can identify weak evidence in someone else's paper but not their own, the problem is peer-response structure and self-assessment, not their conception of revision, and the intervention is different.
2. Restructure the next essay cycle given 45-minute periods and the requirement to finish two more major essays. Be specific about what you cut and what you add.
Read the worked answer
Cut: one timed essay. Four timed drafts produce more of the same; three drafts with real revision produce more learning. Also cut the free-form 15-minute peer session that is generating 'I liked it'-that time is currently buying nothing.
Add:
- A ten-minute modeled revision. He revises one of his own paragraphs live, thinking aloud: 'This evidence doesn't actually support my claim-I need to cut it or change the claim.' Students watch a decision being made, not a correction.
- One lens per round. Round one is evidence only. Round two is organization or counterargument. A single focus makes peer response concrete and makes his feedback consistent with theirs.
- A structured peer protocol: the writer states what they want help with, the reader answers three fixed questions tied to the rubric trait, and comments must quote the text.
- A revision memo: two sentences naming what changed and why. This is what makes the revision visible and gradable.
Assessment alignment: score the targeted trait on the revision, and give process credit for the memo. If he grades everything on every draft, neither he nor the students can tell what improved.
3. How will you know within one cycle whether this worked, and what result would count as failure?
Read the worked answer
Use a proximal measure, not the holistic score. Holistic rubric scores are noisy and unlikely to move in one cycle; judging the change by them would likely produce a false negative and a retreat to the old routine.
Measure:
- Code the revisions. For each draft pair, classify every change as surface or meaning-level. The target is a shift in the proportion of meaning-level changes-organization, evidence, claim, counterargument.
- Score the single targeted trait across draft and revision.
- Check transfer on the next essay's first draft: does evidence use improve without the scaffold?
Failure conditions:
- Meaning-level revisions appear only in class, under his eye, and disappear from at-home drafts.
- Peer comments return to 'add more detail' despite the protocol. That would suggest the protocol is being complied with rather than used, and the next step is modeling peer response, not adding another checklist.
- Revision quality rises but first-draft quality does not. That is not failure-it is the intended sequence-but it means the mini-lessons need to migrate earlier into the drafting phase.