Who Earned the 4? Pair Role-Play and Individual Scores
Ms. Duarte runs a four-minute information-gap role-play in Spanish IV: one partner is a travel agent, the other a client planning a two-week trip on a $2,000 budget. There are 12 pairs and one holistic 1–5 rubric. Reviewing the recordings, she finds that in five pairs one student produces roughly 80% of the turns and drives the transaction, while the quieter partner produces short but accurate, appropriate utterances ("¿Cuánto cuesta?", "Prefiero salir el lunes"). Those quieter students score 2s; their dominant partners score 4s and 5s. In four other pairs, students drop into English for 20–30 seconds whenever they hit a vocabulary gap, then return to Spanish. Ms. Duarte must submit individual grades by Friday.
1. Is the holistic pair score a valid measure of each student's interpersonal ability?
Read the worked answer
- No, as currently applied. The construct is each individual's interpersonal proficiency; a holistic pair score folds in partner effects and turn-taking artifacts that are irrelevant to the construct.
- The evidence contradicts a "weak partner" reading. Those partners produced short, accurate, appropriate turns that accomplished communicative functions. At Intermediate, short turns are developmentally normal; the criterion is function (asking, answering, negotiating), not length.
- The dominant partner is not automatically penalized either. Volume may reflect genuine strength. Score the evidence in the recording, not sympathy for the quieter student - the goal is an accurate individual judgment, not an inversion of the bias.
- Practical fix by Friday: score each speaker separately from the recording against function-based descriptors, and note that one task is thin evidence. Rotate partners on the next task and average across two performances before the grade becomes consequential.
2. How should the 20–30 seconds of English be treated in scoring and in instruction?
Read the worked answer
- Distinguish code-switching as a resource from compensation for a missing strategy. In this task the English stretch is almost certainly compensation: a vocabulary gap, no circumlocution available, then a return to Spanish.
- Set the policy before the task and apply it uniformly. Since the transaction was otherwise completed in Spanish, do not zero the performance; record the episode as evidence of limited strategic competence in that rubric row. Consistency matters more than severity.
- Teach the missing strategy explicitly: Es una cosa que se usa para…, No sé cómo se dice, pero…, ¿Cómo se dice…?, O sea…. Rehearse it once before the assessment so the gap is a performance issue, not a never-taught issue.
- A help card is legitimate scaffolding, not a crutch, at this level - but decide in advance whether it is allowed during the graded task.
3. Redesign the task so that both partners must interact and the rubric captures it.
Read the worked answer
- Split the information asymmetrically: the agent holds the itinerary options and prices; the client holds the constraints plus two non-negotiables (a food allergy, a fixed departure). Neither can finish alone.
- Require per-student deliverables: obtain three pieces of information and provide three (checklist), and each student must ask at least one follow-up question. This makes interaction obligations, not turn share, the thing being scored.
- Shorten to 2.5–3 minutes. Four minutes rewards the dominant speaker and exhausts the Intermediate speaker; add one scripted "breakdown" moment the pair must repair.
- Rubric with separate rows and can-do descriptors: initiates/asks, responds appropriately, repairs a breakdown, sustains the exchange. Score each student on each row from the recording.
- Report to students using the rows, so the quiet partner sees that accurate short turns are real evidence, and the dominant partner sees that dominating is not itself a criterion.