When 68% on Grammar Does Not Mean They Can Order
A German I department with seven sections and four teachers gives a common summative on the accusative case and 'möchten.' The test is 40 fill-in-the-blank items; the department average is 68%. On a separate oral ordering task, only 12 of 28 students in one section could complete a basic request without reading from notes. Many could fill in 'Ich möchte einen Kaffee' but could not respond to 'Mit Milch?' The unit lasts three weeks, and the textbook chapter includes food vocabulary. The district requires an integrated performance assessment (IPA) each unit and common grading. Two teachers want to keep the 40-item test because it predicts quiz scores; one says it measures nothing useful. The department has one 45-minute common assessment window and needs a defensible compromise.
1. Interpret the 68% average alongside the oral result. What does this evidence suggest about the current test's construct validity?
Read the worked answer
The 68% average shows that students can supply forms in isolated sentences, but the oral result shows that many cannot use the same forms to exchange meaning. That is a construct-validity problem: the test samples recognition and controlled production of one structure, not the ability to order food or respond to a follow-up. The two results are not contradictory; they measure different constructs. A student can choose 'einen' correctly on a blank and still fail to retrieve 'Ich möchte einen Kaffee' under communicative pressure. The 68% average may also reflect item difficulty and guessing rather than stable proficiency. The department should not conclude that students 'know the accusative' or that the oral task is unfair. It should conclude that the current summative is underrepresenting interpersonal and presentational use. Keep the grammar information as one data point, but do not treat it as the unit's main outcome.
2. Propose an IPA for this unit with three tasks and the evidence each provides. Keep it within the 45-minute common window plus class time.
Read the worked answer
Interpretive task, 12–15 minutes: students read a short café menu and listen to a 45-second order dialogue, then answer three items: what the customer orders, one follow-up question the server asks, and one detail about the bill. Evidence: comprehension of food vocabulary and the request sequence.
Interpersonal task, 3–4 minutes per pair, done across two class days: a role-play in which one student orders and the other is the server; the server must ask at least one follow-up ('Mit Milch?', 'Zum Hieressen oder zum Mitnehmen?'), and the customer must respond. Evidence: spontaneous use of 'möchten,' accusative forms, and repair strategies. Record a short sample for moderation.
Presentational task, 15–18 minutes in the common window: students write a brief café order or a short review using a model, with a required greeting, two items ordered, one question, and a closing. Evidence: controlled but meaningful production.
The 45-minute window can hold the interpretive and presentational tasks; the interpersonal task runs in class. Grammar is embedded in all three tasks rather than tested as a separate 40-item section.
3. How can the department reach common grading and bring the reluctant teachers on board without discarding their concern about foundational grammar?
Read the worked answer
Start by agreeing on what the department wants students to be able to do, then show where the current test and the IPA each provide evidence. Keep a short, embedded grammar check-perhaps eight to ten items drawn from the IPA context-as one data source, not as the summative. This respects the reluctant teachers' concern that form matters while removing the 40-item test from its dominant role.
Co-score three anchor samples together: one strong, one typical, one weak. Use a shared rubric with two dimensions, communicative task completion and language control, and calibrate until raters agree within one level. Pilot the IPA in one section and compare results with the old test; bring the data to the next department meeting rather than arguing from principle. Assign one teacher to adapt the interpretive task and another to write the interpersonal scenario, so ownership is shared. Set a review date after the first unit. The compromise is not 'no grammar'; it is 'grammar as part of communication, with separate evidence for form when needed.'