The problem with assessing speaking is one that frustrates teachers in every context: it is simultaneously the skill students most want to develop and the hardest skill to evaluate fairly. A student who writes beautifully can produce evidence you can mark at your desk on Sunday afternoon. A student who speaks fluently and persuasively leaves no permanent record. The assessment happens in real time, under time pressure, while you're simultaneously managing the class.
The result is that speaking assessment is often done poorly - too impressionistic, too infrequent, or too detached from actual classroom performance. This guide offers practical frameworks that work in real classrooms without requiring one-to-one testing time you don't have.
YapYapGo is a classroom speaking practice tool for ESL and EFL teachers. Its structured pair work modes create consistent, repeated speaking opportunities that are far more assessable than unstructured conversation. Here's how to build a workable assessment system around them.The core problem with speaking assessment
Teacher-assessed one-to-one oral tests have two problems. First, they're logistically impossible to run frequently in a class of 25+. Second, they don't reflect how students actually communicate - a one-to-one test with a teacher is a very different context from pair discussion with a peer.
A better approach uses three complementary methods:
- Observation during pair work (formative, frequent, low-stakes)
- Peer assessment (formative, teaches students to analyse language)
- Structured performance tasks (summative, less frequent, higher stakes)
No single method is sufficient. Together, they give a complete picture.
Simple rubric frameworks
The 4-criteria rubric
For general speaking assessment, four criteria cover most of what matters. Rate each on a 1-5 scale:
Fluency: Does the student speak without excessive pauses? Do they keep the conversation moving? (1 = very hesitant; 5 = fluent and continuous) Coherence: Does the student organise their ideas clearly? Do they use connectives and discourse markers? (1 = disconnected; 5 = well-organised with clear logic) Vocabulary: Does the student use a range of appropriate words? Do they avoid repeating the same words? (1 = very limited; 5 = wide, flexible range) Communication: Does the student succeed in communicating their meaning? Does their partner understand them? (1 = frequent breakdown; 5 = always understood)Note what is absent: grammar accuracy is not a criterion. This is intentional. Grammar can be assessed in writing. In speaking assessment, fluency and communicative success are more important. If you want to include grammar, replace "communication" with "grammatical range."
The quick three-band descriptor
For very fast formative assessment during class, a simpler three-band system:
- Developing: Communicates with effort, frequent pauses, limited vocabulary range, mostly understood
- Achieving: Communicates clearly, some hesitation, adequate vocabulary, always understood
- Excelling: Communicates fluently, wide vocabulary, discourse management, easy to follow
Write the level next to each student's name during pair work rounds. After four or five observations across different activities, a clear pattern emerges.
The IELTS-aligned rubric (B1-C1)
For students working towards IELTS or other proficiency exams, align your classroom rubric to the exam criteria:
| Criterion | Band 5 | Band 6 | Band 7 | Band 8 |
|-----------|--------|--------|--------|--------|
| Fluency | Long pauses, repetition | Some hesitation | Mostly fluent, self-corrects | Natural fluency |
| Vocabulary | Limited, repetitive | Adequate range | Some less-common words | Flexible, precise |
| Grammar | Simple structures | Some complex, errors | Complex, mostly accurate | Wide range, minor errors |
| Pronunciation | Effort required | Mostly clear | Clear, some features | Natural, easy |
This serves double duty: it assesses the student and teaches them the criteria they'll be judged by in the exam.
Tool tip: YapYapGo's IELTS mode creates exam-format speaking conditions with Parts 1, 2, and 3 in sequence. Running formative assessment during these sessions using the IELTS rubric gives students realistic band estimates that directly inform exam preparation. A classroom countdown timer mirrors real exam timing during assessment sessions.
Practical observation systems
The clipboard method
Carry a clipboard with your class list and a simple grid. During each pair work round, listen to two or three pairs for 60-90 seconds each and mark a quick assessment using your chosen rubric. Across a lesson, you observe 8-10 students. Across a month, you build a complete picture of the class.
Key: note specific examples, not just grades. "Used 'on the other hand' naturally - B2 coherence marker" is more useful than "7/10."
The focus observation
Rather than trying to assess everything every lesson, focus on one criterion per lesson. This week, observe fluency only. Next week, vocabulary. This produces better quality observations than trying to track all four criteria simultaneously.
Brief students: "Today I'll be listening specifically for how you use connectives." This raises their conscious attention on that feature, which itself is a pedagogical benefit.
The peer observation card
For every third or fourth speaking activity, give students a peer observation card with one specific focus: "Mark one tick every time your partner uses a linker (because, however, for example, etc.)." They track this while listening; after the activity, they share their card and discuss.
This generates self-assessment data for the student and saves your observation time.
Summative assessment without one-to-one tests
The prepared presentation
Students prepare a 2-3 minute talk on an assigned or chosen topic outside class. They deliver it in class to a small group (3-4 students) who complete a simple peer observation form. You circulate and observe multiple groups. Each student is assessed on their actual delivery, not their notes.
The group format means you can observe 6-8 students simultaneously rather than one at a time.
The paired performance task
Two students are given a task together - a negotiation scenario, an information gap activity, a problem-solving task. They complete it while you observe one of them specifically. After the task, brief whole-class discussion. You've assessed one student in context without stopping the lesson.
Rotate which student you observe across the class over several sessions.
The portfolio approach
Students keep a simple record of their speaking over a term: what activities they completed, a self-assessment after each, and three "evidence moments" where they note something they said that represented their best language use. You review portfolios at the end of the term rather than conducting oral tests.
The portfolio develops metacognitive awareness while reducing your assessment workload. See our post on peer assessment for speaking for how to build peer assessment into this portfolio approach.
A random student picker and activity timer are both useful during summative tasks - the picker for fair student selection, the timer for keeping assessment sessions consistent in length. For formative observation during pair work, YapYapGo frees you from logistics so your attention can go entirely to assessment.
Sources:
- Luoma, S. (2004). Assessing Speaking. Cambridge University Press. - The comprehensive framework for speaking assessment in ELT.
- McNamara, T. (1996). Measuring Second Language Performance. Longman. - Validity and reliability in oral language testing.
- Sato, M. & Lyster, R. (2012). Peer Interaction and Corrective Feedback. Studies in Second Language Acquisition. - Peer observation as a valid assessment tool.