IELTS Speaking Band Scores: Criteria, Descriptors, and How to Improve
Your IELTS Speaking score is the average of four equally weighted criteria, each marked on a band scale from 1 to 9. Understanding exactly what each criterion measures — and what distinguishes one band level from the next — is the most direct path to a targeted score improvement. Studying harder without knowing which criterion is holding you back is the most common cause of stagnation.
According to IDP Education’s published test statistics, the global average IELTS Speaking score in 2023 was 6.6. The most frequently reported bottleneck for candidates sitting in the 6.0 to 6.5 range is the Fluency & Coherence criterion — not, as many assume, grammar or vocabulary. Once you have identified your weak criterion here, the IELTS Speaking tips guide gives you targeted strategies for addressing each one before test day.
The Four Assessment Criteria
| Criterion | Weight | What It Measures |
|---|---|---|
| Fluency & Coherence | 25% | Smoothness of delivery, logical organisation of ideas, appropriate use of connective language, and the ability to sustain extended speech without unnatural hesitation |
| Lexical Resource | 25% | Range and precision of vocabulary, collocational accuracy, use of idioms and less common items, and ability to paraphrase effectively when precise vocabulary is unavailable |
| Grammatical Range & Accuracy | 25% | Variety of grammatical structures used, complexity of sentence construction, and the frequency and significance of grammatical errors |
| Pronunciation | 25% | Clarity and intelligibility, natural use of stress and rhythm, intonation patterns, and the ability to be understood without significant listener effort |
Fluency and Coherence
Fluency is not about speaking quickly. The band descriptors define fluency as the ability to speak at length “without noticeable effort or loss of coherence.” A candidate who speaks deliberately and slowly but connects ideas clearly scores higher in this criterion than a candidate who speaks rapidly but with frequent self-correction and topic drift.
Coherence — the logical organisation of ideas — is the more neglected half of this criterion. Examiners listen for topic sentences that establish direction, discourse markers that signal logical relationships, and endings that resolve the point rather than trailing off. A response that is technically fluent but lacks internal logic fails coherence even if individual sentences are well-formed.
Lexical Resource
The Lexical Resource criterion rewards breadth, precision, and naturalness. The band descriptors specifically penalise “limited vocabulary that prevents discussion of some topics” at band 5, and reward “idiomatic vocabulary with only occasional inaccuracies” at band 8. For a topic-by-topic vocabulary bank covering the most common IELTS Speaking themes, see the Speaking vocabulary guide. Candidates aiming for band 7 and above will also benefit from the formal academic terms in the IELTS Academic Word List, which covers the higher-register vocabulary that distinguishes band 7 Part 3 answers from band 6 ones.
The key insight is that examiners reward attemptsat sophisticated vocabulary even when imperfect. A candidate who uses “exacerbate” incorrectly demonstrates a more ambitious lexical range than a candidate who avoids it by always saying “make worse.” The marking guidance is explicit: minor lexical errors in ambitious vocabulary are penalised less than chronic restriction to a narrow safe vocabulary.
Grammatical Range and Accuracy
This criterion assesses two distinct things: the range of grammatical structures you use, and the accuracy with which you use them. Both matter independently. A candidate who produces perfectly accurate simple sentences throughout the test will not exceed band 6 in this criterion, because range is explicitly required for band 7 and above.
Useful complex structures for speaking include: conditional clauses (“If I had known earlier, I would have approached it differently”), relative clauses (“The approach that has proved most effective is…”), passive constructions (“It has been argued by several researchers that…”), and reporting verbs (“Studies suggest / indicate / demonstrate that…”). These structures do not require elaborate vocabulary — they require structural confidence. They are the same structures that distinguish high-scoring band 9 Writing Task 2 responses from average ones, making them worth studying in both contexts.
Pronunciation
Pronunciation at band 7 and above requires clear intelligibility and natural use of prosodic features — stress, rhythm, and intonation — not a specific accent. The band descriptors are unambiguous that a regional or non-native accent is not penalised, provided it does not impede understanding.
The most common pronunciation errors that do cost marks are: consistent word stress errors (placing stress on the wrong syllable), failure to use sentence stress to convey meaning, and monotone delivery that makes it difficult for the examiner to track which information is new or important. These are all teachable and correctable through deliberate practice.
Band Level Comparison Table
| Band | Fluency & Coherence | Lexical Resource | Grammatical Range & Accuracy | Pronunciation |
|---|---|---|---|---|
| 5 | Speaks with obvious pauses; limited coherence | Limited vocabulary; frequent repetition | Uses basic sentence forms; many errors | Often unintelligible; L1 features frequently intrude |
| 6 | Willing to speak at length; some hesitation and loss of coherence | Some flexible vocabulary use; some errors in word choice | Mix of simple and complex structures; errors in complex forms | Generally intelligible; L1 accent does not impede communication |
| 7 | Speaks at length with minimal effort; some hesitation | Flexible use of vocabulary; some inaccuracies in idiom | Frequent complex structures; some errors that do not impede | Easy to understand; some features of L1 remain |
| 8 | Fluent with only occasional repetition; coherent throughout | Wide vocabulary used naturally; rare inaccuracies | Wide range of structures; most are accurate; minor slips only | Easy to understand; uses stress and intonation to convey meaning |
| 9 | Speaks fluently with no apparent effort; coherent and logical | Full command of vocabulary; precise and idiomatic | Full grammatical control; structures are flexible and accurate | Effortlessly intelligible; all prosodic features used effectively |
Band 6 vs. Band 8: A Direct Comparison
Seeing the same question answered at two different band levels makes the distinction concrete in a way that descriptor tables alone cannot.
Question:“Do you think governments should do more to promote healthy eating?”
Band 6 answer:“Yes, I think governments should do more. Many people eat unhealthy food because it is cheap and easy to find. If the government makes healthy food cheaper, more people will eat it. Also, they can put education in schools about nutrition. I think this is important for the health of the country.”
Band 8 answer:“I’d argue that governments have a clear public health mandate here — dietary disease costs healthcare systems an enormous amount, so there’s a straightforward economic case for intervention, quite apart from the ethical one. That said, the most effective interventions are probably structural rather than educational: food labelling reform, sugar taxes, and zoning restrictions on fast food near schools have all shown measurable effects in countries that have implemented them. Pure awareness campaigns, by contrast, tend to reach people who are already motivated to change. What I find interesting is that the most resistance typically comes not from the public but from the food industry, which suggests that the barriers are political rather than practical.”
The band 8 answer is longer, uses a wider range of vocabulary (“mandate,” “intervention,” “zoning restrictions,” “measurable effects”), employs complex sentence structures (relative clauses, passive constructions), and demonstrates a more sophisticated analytical structure — problem, mechanism, evidence, nuance.
Self-Assessment: Diagnosing Your Weak Criterion
Recording yourself answering five Part 3 questions and listening back critically takes approximately 30 minutes and provides more diagnostic information than any practice test. Use this four-question checklist:
- Fluency & Coherence: Do your answers flow without unnatural pauses? Can you follow the logical thread from one sentence to the next without losing it? If not, practise discourse markers and PEEL structure.
- Lexical Resource:Are you using the same ten words repeatedly? Count the number of different vocabulary items per 100 words. If it’s low, build topic-specific collocation banks and use two new items in every practice session.
- Grammatical Range: Are all your sentences simple or compound? If you cannot identify at least one relative clause, one conditional, and one passive construction in your recording, your range is limiting your score. Practise embedding these structures into familiar answer templates.
- Pronunciation: Can a listener follow every sentence without effort? Are you placing stress on the right syllable in multi-syllable words? If you are uncertain, use a text-to-speech tool to check stress patterns on words you use frequently.
Isolating and targeting one criterion per week of focused practice is more effective than attempting to improve all four simultaneously. Most candidates scoring at band 6 to 6.5 have one criterion that is significantly lower than the others — addressing that single criterion has a disproportionate impact on the overall score. To put this self-diagnosis into action, follow the structured daily routine in the IELTS Speaking practice guide.