Placement Decisions with a 25-Point Standard Error
Maya is a rising ninth grader entering a district German program. Placement uses a 60-item multiple-choice test of vocabulary and grammar plus a 10-minute writing sample. Maya's raw score is 62%, which the conversion table maps to a scale score of 480. The German II cut score is 500, so the building placement officer has assigned her to German I. The scale score has a standard error of measurement of 25 points. Her writing sample, scored holistically by two teachers, is 90 words in 10 minutes with varied vocabulary, a clear past-tense narrative, and few errors other than two case endings. The department is also debating moving to a proficiency-based sequence with an interpretive and interpersonal component, because the current test has no listening or speaking.
1. Evaluate the placement decision statistically and recommend a placement, showing your reasoning.
Read the worked answer
A scale score of 480 is 20 points below the cut score of 500, and the standard error of measurement is 25. That difference is less than one standard error. The approximate 68% confidence interval is 455–505, and the 95% interval is roughly 430–530. Her observed score is therefore statistically indistinguishable from the cut score, and placing her by the point estimate alone is not defensible. The writing sample adds independent evidence: 90 words in 10 minutes with a clear past-tense narrative is above what most German I entrants produce. The defensible procedure is to collect one more piece of evidence, a 10-minute oral interview and a short listening task, then place her in German II with a support plan if the interview confirms the writing profile. If the interview reveals weak listening, German I with an acceleration option is reasonable. The key point is that the cut score should not be applied mechanically when the measurement error is larger than the distance to the cut.
2. What is the strongest argument for adding interpretive and interpersonal tasks, and what would you need to do to keep placement fair if you add them?
Read the worked answer
The current test measures recognition of vocabulary and grammar, not proficiency. It cannot observe whether a student can understand spoken German or sustain a conversation, which are the abilities that German II actually requires. Adding a short authentic listening task and a five-minute recorded conversation would align the test with the construct the placement decision claims to measure. Fairness, however, requires more than adding tasks. Cut scores should be set through a standard-setting procedure with trained judges who review performances at the boundary, not by carrying over the old 500. Raters need a common rubric and calibration. The department should check whether items function differently for heritage speakers and students with prior travel experience, because a grammar test can underplace exactly those students. Finally, report confidence intervals alongside placements and allow a review pathway when a score falls within one standard error of the cut.
3. Maya's writing sample has two case-ending errors. How would you use that information without overreacting?
Read the worked answer
Two case-ending errors in 90 words is a minor pattern, not a reason to place her in German I. At this level, the sample shows fluent production, a controlled past tense, and varied vocabulary; the case errors are likely the residue of a system still being consolidated. Use them diagnostically. If she enters German II, give a short ungraded diagnostic on accusative and dative forms in the first week, then a focused mini-lesson and periodic monitoring rather than a full remedial unit. Do not make case endings the lens through which the whole sample is judged, and do not treat a small error count as evidence that she cannot handle German II. The placement decision should rest on the preponderance of evidence - the statistically inconclusive test score, the strong writing sample, and the oral interview - with the case endings noted as a targeted instructional objective.