Eighty Percent Comprehension, Thirty-Eight Percent Revision

Split the data before deciding.

• 80% on comprehension says the passages and reading level are accessible - this is not a fluency problem.
• 38% correct on revision says the assessed skill, not the text, is the gap.
• The shortest, simplest distractor winning is a test-taking heuristic, not a knowledge error: students are choosing the option that looks easiest to read.
• "Didn't make sense" plus guessing suggests they lack a procedure for judging a sentence and default to sound.

Decision: proceed with the new unit rather than halting it, and embed a two-day revision mini-unit into the first days - grammar in isolation for a week would waste time and would not transfer to choosing among options. The unit's texts can carry the revision work.

Set a decision rule now: if a short revision probe at the end of day two still shows fewer than half the class choosing well-justified revisions, add a third day and pull the guessing students (about five) for small-group work.

What not to do: conclude the students "can't write," delay the whole unit, or wait for the benchmark to confirm what the interim already showed.

Day 1 - mentor sentences and sentence combining (no test format).

• Four pairs of sentences; students combine them and compare two valid combinations, arguing which is clearer.
• Establish the criteria students will use all year: complete, clear, and consistent in structure.
• Exit task: combine two sentences two different ways and mark which is better and why.

Day 2 - applying the criteria to options.

• Present one sentence with four revisions. Students rank them and must eliminate two with a written reason.
• Explicitly name the trap: "The shortest option is often the one that drops information. Check what each choice keeps and loses."
• Strategy card students keep: (1) read the original, (2) list what it must keep, (3) test each option against the list, (4) never choose on length alone.

Formative check (10 minutes, end of day 2):

• Three items: one revision choice with a written elimination reason, one "which two are both correct and how do they differ," one open revision of a run-on.
• Decision rule: if fewer than half justify their eliminations, add day 3 and pull the guessing students for guided practice. If the class is above 75% with reasons, move into the unit and revisit in a warm-up each week.

Watch for: students who choose the right answer for the wrong reason. The written reason is the actual assessment.

Do not administer the copied test as written. Thirty comprehension items assess the skill the class already passed at 80% and assess nothing about the skill they failed. It also produces a grade that will look fine and teach the team nothing.

Practical fix, given copies already exist: rebuild the test on the same paper if the front office can copy by Friday; if not, use the comprehension items as a short warm-up and administer a revised assessment the following Tuesday, telling students plainly why. Reusing the printed sheets as scratch or homework is a legitimate way to absorb the cost without pretending they are the assessment.

Blueprint for the rebuilt test (roughly 12–15 items, not 30):

• Two or three comprehension items as a baseline.
• Six to eight revision items requiring elimination or justification.
• Two constructed responses: revise a paragraph and explain one choice.
• Include the trap distractor deliberately and check afterward who chose it - that is the item that tells her whether day 2 landed.

Going forward: write items against the standard, not the textbook section; run item analysis after every common assessment (percent correct, distractor counts, who guessed); and share the blueprint with the team so the district benchmark week is spent on data, not on discovering a mismatch. Bring the distractor pattern to the next PLC meeting - if other teachers see it too, the fix belongs in the grade-level plan, not one classroom.