Assessing Speaking in a Noisy Class of 32
A teacher of a class of 32 learners aged 10–11 (A2) has 40 minutes and one assistant for 15 of those minutes. She must report each learner's speaking progress to parents and decide what to reteach next week. She is considering: a whole-class oral quiz, a pair information-gap task with a three-point checklist, a two-minute individual interview with six learners, or a group poster presentation. The room is echoey, learners are noisy, and she can record on her phone. The parents' report needs a simple secure or working-towards judgement per learner.
1. Select the main assessment activity and justify it against validity, reliability and practicality. Explain why at least one alternative is unsuitable.
Read the worked answer
Choose the pair information-gap task with a three-point checklist as the main activity, plus a short individual interview sample.
- Validity: the reporting decision is about spoken interaction - asking, clarifying, taking turns. A pair information-gap task elicits that directly. A whole-class oral quiz mostly elicits choral answers and tells you little about individual speaking.
- Reliability: a three-point checklist (for example: asks for repetition; gives clear instructions; uses two target phrases) gives the same criteria for every learner. Recording six pairs on the phone allows moderation and a second listen.
- Practicality: 32 learners in 40 minutes with one assistant for 15 minutes makes individual interviews for all impossible. Pair tasks keep the class working while the teacher and assistant observe.
- Why not the poster presentation: it confounds speaking with writing, drawing and group dynamics, and a group product cannot support an individual secure or working-towards judgement.
The individual interviews are a sample, not the main event: six learners across the range calibrate the checklist and catch learners who hide in pairs.
2. How would you turn the pair task into evidence for a secure or working-towards judgement for 32 learners?
Read the worked answer
Turn the task into evidence in four steps:
- Define three observable criteria and write them as descriptors, not scores: initiates an information question, responds to a clarification request, uses the target phrase accurately enough to be understood.
- Observe in rotations. The assistant takes eight pairs for 15 minutes while the teacher takes eight; the teacher then samples the remaining pairs during the task. Each pair is observed once, not judged from a single utterance.
- Record a sample of six pairs, two from each attainment band, and listen again after class to moderate the secure boundary.
- Use a best-fit judgement across two tasks: the pair task plus one short teacher-led question in the next lesson. A learner is secure only if the evidence appears in both, not from one lucky turn.
Avoid: judging by volume of talk, by confidence, or by one recorded pair. The checklist must be about the interaction, not about personality.
3. What will this assessment not tell you, and how would you cover that gap?
Read the worked answer
What it will not tell you: written accuracy, sustained monologue, reading or listening progress, and how the learner performs with an unfamiliar interlocutor. It also under-samples quiet learners who let a partner do the talking, and it can be distorted by pairing (a strong partner scaffolds or dominates).
Cover the gaps:
- Add a five-minute written exit task (three sentences using the target phrases) to check form.
- In the next lesson, take a short whole-class observation of two learners doing a monologue-style instruction task.
- Pair learners deliberately across the task and note who initiates; if a learner never initiates, re-pair and re-observe rather than inferring.
- For the parents' report, state the evidence base: 'secure in a supported pair task' is more honest than a blanket secure.