The Unit Test That Measures the Wrong Thing
Mr. Patel teaches eighth-grade U.S. history. His unit test on westward expansion has ten multiple-choice items with long stems and two essay prompts, including: 'Analyze the extent to which manifest destiny represented both opportunity and hypocrisy, citing at least three pieces of evidence.' Twelve of his 96 students are English learners, spanning WIDA Levels 1 through 4. On this test, ELs averaged 52% and non-ELs averaged 81%. One Level 2 student wrote a paragraph in Spanish explaining that settlers took land from Native peoples, drew and labeled a diagram showing the movement of settlers and the displacement of nations, and received 2 out of 20 on the essay. Mr. Patel concludes that the ELs 'didn't study' and proposes after-school test prep plus retention for three of them.
1. Diagnose the validity problems in this assessment. Which problems relate to language and which relate to the history content?
Read the worked answer
The central problem is that the test measures history plus academic English and then reports a single number. 'Analyze the extent to which' is an abstract argument structure that many students, EL or not, have never been taught to execute; 'hypocrisy' is a low-frequency evaluative word that can block a reader before any history is shown. For a Level 2 student, decoding and interpreting the prompt may consume the task before content enters the picture. The long multiple-choice stems likely carry the same problem, so the 52% may understate what ELs know.
That is construct-irrelevant variance: written English is not the target of a history standard. The Level 2 student's Spanish paragraph and labeled diagram show accurate historical understanding of displacement and settlement, and the format offered no route to credit - that is a scoring problem, not a knowledge problem.
What does relate to content: whether the essay requires evidence the students actually had access to, whether the unit taught the reasoning structure being tested, and whether the prompt is answerable at the intended level of specificity. The gap between EL and non-EL means is not evidence of effort; it is a signal to inspect the instrument.
2. Redesign the assessment so English learners can demonstrate the same historical content standard. Give specific moves.
Read the worked answer
Separate the content standard from the language of the test. Keep the historical thinking goal - explaining competing motives, opportunities, and consequences of westward expansion - and build multiple routes to it: a prompt written in shorter sentences with one question per sentence and a glossary; a document-based task with visual primary sources where students annotate before writing; an oral interview or recorded explanation scored with a short rubric; and a choice board including a timeline with captions, a labeled map with brief explanations, or a paragraph.
Pre-teach the task language ahead of time ('extent,' 'motives,' 'consequences') and supply sentence frames for framing a claim and citing evidence. Permit a bilingual word bank and L1 for planning, with the final product in English or a mix, depending on the student's level. Score the historical claim, use of evidence, and reasoning on a content rubric, and score language conventions separately so errors are not counted twice. Give ELs the same preparation as peers: model a strong and a weak response and rehearse the reasoning routine in pairs. Rigor stays in the reasoning; only the linguistic barrier changes.
3. What should Mr. Patel do about grading and the retention proposal?
Read the worked answer
Grade content and language separately. The history grade should reflect the accuracy and support of the historical claim; spelling, articles, and sentence complexity belong on a language rubric tied to each student's English language development goals, not folded into the content grade. The Level 2 student's Spanish paragraph plus diagram is valid evidence of the content standard once the task permits it and should be credited.
The retention proposal is not supported by the data. The scores reflect an instrument that added a language barrier, so they cannot be read as effort or as failure to learn. Retention generally does not improve achievement and is associated with worse outcomes, especially for English learners, who also lose access to age-appropriate content and peers. Before any retention discussion, Mr. Patel should revise the assessment, re-teach the language of the task, check whether ELs had the same access to preparation and materials, and gather multiple sources of evidence across the unit. After-school test prep is reasonable only if it targets the language of the assessment alongside the content - and only after the instrument itself has been fixed.