Evidence Without Explanation in Seventh-Grade Argument Writing
Your seventh-grade ELA team has just finished a three-week argument-writing unit built around two short op-eds on school start times. You have 27 students and a departmental rubric with four criteria: claim, evidence, explanation, and conventions. Scoring the on-demand essays, you find that 19 of 27 students quote or paraphrase relevant evidence, but only 6 explain how that evidence supports the claim. The pattern is consistent: a typical paper reads, 'The author states that teens need more sleep. Also, mornings are hard for teenagers.' One student, Devon, does explain his evidence but only after you prompt him orally; his written draft contains none of what he said. The pacing calendar says the narrative unit starts in eight days. A colleague proposes swapping two days for multiple-choice practice on identifying claims and evidence, and the principal wants your data notebook updated before the next team meeting.
1. What is the most probable explanation for the pattern in these essays, and how should you spend the eight days before the narrative unit begins?
Read the worked answer
Diagnosis. The data do not show a problem with finding evidence, since 19 of 27 students select relevant quotations. They also do not show a problem identifying a claim, which is what the proposed multiple-choice practice would address. The gap is the explanatory move: students do not yet know that an explanation must name the logical link between the evidence and the claim. Devon is the strongest evidence for this reading. His reasoning exists, but it lives in speech, not on paper, so this is a production problem rather than a comprehension problem.
Plan for the eight days.
- Model the missing move with a think-aloud, placing a 'quote-and-run' paragraph beside an elaborated one and naming exactly what the second does.
- Give a two-sentence elaboration frame ('This matters because...'), used deliberately and then faded.
- Have students revise the same essay instead of drafting a new one. A new prompt spends working memory on inventing content; revision keeps attention on the target move, and one revision cycle fits inside eight days.
- Use short exit tickets containing only a claim, a quotation, and a two-sentence explanation, scored against three look-fors.
The multiple-choice request targets a skill the data say is already present. A better compromise is a five-minute retrieval warm-up on claim identification while the bulk of time goes to writing.
2. How would you respond to the typical paper and to Devon's draft, and what formative check would tell you the instruction worked?
Read the worked answer
Typical paper. Do not mark every instance. Choose one anonymous sample, project it, underline the quotation, and ask the class, 'What does this quotation prove?' Write the class's answer as a sentence directly beneath the quotation. This makes the invisible move visible and gives 19 students a concrete target rather than a correction.
Devon. Confer briefly, transcribe his oral explanation onto a sticky note, and have him compare it with his draft. The point is to show him that he already has the reasoning and only needs to write it. Then have him insert and extend it. Do not simply tell him to 'explain more' - that instruction assumes he knows what explaining looks like.
Feedback language. Use the rubric's own words: Explanation - states why the evidence supports the claim. If students cannot connect feedback to a criterion, the rubric is decoration.
Formative check. Give a new claim and a short quotation, five minutes, one paragraph. Score three look-fors: (a) the quotation is relevant, (b) a sentence explicitly links evidence to claim, (c) the link is specific rather than 'this shows it is true.'
Decision rule. If fewer than about 70 percent of students produce a linking sentence, extend by one day and pull the rest for small-group work. If they can link but not elaborate, the problem is depth, not structure, and the next mini-lesson should target elaboration.
3. How do you handle your colleague's test-prep proposal and the pacing calendar without simply refusing?
Read the worked answer
Treat the proposal as a question about which skill is being assessed, not as a loyalty test.
- Separate the two skills. Recognizing a claim in a multiple-choice item and producing an explanation in prose are different performances. Our data show recognition is largely in place and production is not, so time should follow the deficit.
- Offer a bounded compromise. Accept one ten-minute retrieval set on claim and evidence identification as a warm-up, and protect the remaining time for writing and revision. This honors the colleague's concern about test formats without displacing the instruction the data call for.
- Negotiate the calendar with evidence, not preference. Bring the scored baseline, the exit-ticket results, and a one-page plan showing that the narrative unit starts two days late but that the argument unit's rubric scores will be usable as a pre/post measure. Teams generally accept a documented trade-off more readily than a refusal.
- Update the principal's data notebook with the same numbers so the story is consistent across the team.
What would change my mind: if a short multiple-choice probe showed many students could not identify a claim at all, then recognition would need reteaching first. The decision should stay tied to what the next probe shows, not to the original plan.