The Rubric That Didn't Match the Test
Ms. Whitfield's 8th-grade ELA team spent the summer building a writing rubric with four criteria - organization, evidence, elaboration, and conventions - scored 1–4 each. She uses it faithfully on every essay, and her students' scores on those essays are strong: class average 3.2 out of 4. But on the winter district benchmark, which uses a state-style holistic 1–6 rubric focused heavily on evidence and reasoning, her students score far lower than her classroom data predicted, and several strong classroom writers land in the bottom half. A colleague suggests she 'just teach to the test.' Ms. Whitfield is uneasy about that advice but also worried her rubric is misleading her.
1. Diagnose why Ms. Whitfield's classroom scores and the benchmark scores diverge.
Read the worked answer
The likely cause is a misalignment between the classroom rubric and the benchmark's construct and scale, not that her students suddenly forgot how to write.
Key differences:
- Construct coverage: Her rubric spreads weight across four criteria, including conventions and organization. The state-style rubric weights evidence and reasoning most heavily. A student who is organized and polished but thin on evidence can score 3+ in her system and low on the benchmark.
- Scale and aggregation: Four 1–4 criteria averaged is not equivalent to a holistic 1–6 anchored to different descriptors. A '3.2 average' may reflect strength in criteria the benchmark barely counts.
- Score meaning: Her rubric may be measuring compliance with her instruction rather than the target construct.
This is a validity problem: the classroom assessment is not measuring the same thing as the outcome it is supposed to predict.
2. Evaluate the colleague's advice to 'teach to the test.' What is the more defensible response?
Read the worked answer
'Teach to the test' is a false solution when it means drilling format or narrowing instruction to tested items - that can raise scores without building the underlying skill and may harm transfer.
The defensible response is construct alignment: ensure her instruction and rubric target the same skills the benchmark values - especially evidence and reasoning - while keeping authentic writing tasks.
Concretely:
- Revise the rubric so evidence and reasoning carry appropriate weight, and add descriptors that mirror the benchmark's language for those traits.
- Use the state-style rubric alongside hers on at least some assignments so students see both.
- Analyze a few benchmark essays against her rubric to find where the two systems disagree, then adjust instruction.
This is not teaching to the test; it is making sure the classroom assessment measures the intended learning target, which is the essence of assessment validity.
3. What should Ms. Whitfield do with the winter benchmark data beyond changing her rubric?
Read the worked answer
She should use the data diagnostically, not just as a verdict:
- Disaggregate by rubric criterion to see whether the gap is concentrated in evidence/reasoning - likely, given the pattern.
- Look at the specific essays of students whose scores diverged most; the discrepancy is where the most information lives.
- Bring the pattern to her PLC/team so the rubric revision is a shared decision rather than a solo fix, since the team built the original rubric.
- Adjust instruction for the remainder of the year and re-check with a common assessment.
The principle: assessment data should drive instructional and assessment revisions, and a discrepancy between two measures is a signal to investigate alignment, not to blame students or abandon the classroom rubric wholesale.