Blueprinting a Unit Test Against a Performance Task
The 10th-grade ELA team's common unit assessment for a short-story unit has 40 selected-response items. A quick audit shows 26 items ask students to identify literary terms, authors, or recall plot details; 9 ask for inference; 5 ask students to analyze a passage. The team's average score is 82%, and teachers report the test "looks rigorous" because it is long. On the district's performance task, which asks students to write an evidence-based analysis of an unseen story, only 41% scored proficient. The principal wants assessment data that can guide instruction, and the team has a 45-minute PLC meeting before the next unit. One teacher says the test is fine because it prepares students for the multiple-choice state test; another wants to drop selected-response entirely. The state test includes both selected-response and an essay, and the district's pacing guide allots 6 days for the next unit. The team also has a bank of released passage-based items it has never analyzed.
1. Diagnose the misalignment between the unit test and the district performance task using an assessment blueprint. What categories and cognitive levels are over- and under-represented?
Read the worked answer
The test over-represents recall and recognition: 26 of 40 items ask for literary terms, authors, or plot details; 9 ask for inference; only 5 ask for analysis of a passage. That is roughly 65% recall, 22% inference, and 13% analysis. The district performance task requires evidence-based analysis of an unseen story, which demands transfer, selection of textual evidence, and explanation. The 82% average versus 41% proficient gap is predictable: students are being tested on what was drilled, not on the transfer task. The test "looks rigorous" because of length, but length is not cognitive demand. A blueprint should map items to standards and Depth of Knowledge: recall (DOK 1), inference (DOK 2), analysis and evidence (DOK 3). The team should not drop selected-response entirely, since the state test uses it; instead, make many selected-response items passage-based and evidence-linked. Under-represented: analysis of unseen text, constructed response, and explanation of evidence.
2. Revise the assessment plan: specify a balanced blueprint for the next unit and describe two item types that would generate instructionally useful data.
Read the worked answer
Blueprint for a 40-point unit assessment: 12 recall or identification items (down from 26), 16 passage-based inference items, 8 analysis items using a short unseen passage, and 4 points for a constructed-response paragraph with a two-point evidence rubric and a two-point reasoning rubric. Alternatively, 30 selected-response and one short paragraph. Two useful item types: (1) evidence-linked selected response-"Which sentence best supports the claim that the narrator is unreliable?" with distractors drawn from plausible but irrelevant lines; (2) a two-paragraph constructed response asking students to compare how two passages use a symbol, scored with a claim-evidence-reasoning rubric. These generate data at the item level: which students can select evidence but not explain it, and which can infer but not organize a response. The team can then reteach the specific move rather than the whole story. Keep the state-test format in mind, but do not let format dictate cognitive level.
3. How should the team use the 45-minute PLC and the released item bank to make this change sustainable?
Read the worked answer
Use the PLC as a working session, not a complaint session. First 10 minutes: sort 12 released items by DOK and standard, using a simple t-chart; this builds shared language about cognitive demand. Next 20 minutes: in pairs, adapt or write six passage-based items for the next unit, each with a target standard and an evidence link. Last 15 minutes: agree on a scoring guide for the constructed response and assign roles-who drafts the blueprint, who formats, who brings anchor papers next time. Keep the item bank in a shared folder with tags for standard, DOK, and passage. Sustainability requires a cycle: before each unit, 20 minutes of item review; after each unit, 15 minutes of item analysis (which distractors worked, which constructed responses clustered at which score). The 82%/41% gap should be rechecked with the next performance task. This uses existing time and materials rather than adding a new initiative.