Correcting one of the most common and damaging myths in IELTS prep, straight from the official band descriptors.
If you're Indian, Chinese, Nigerian, Brazilian, Vietnamese, Arabic-speaking, or any of the hundred other first languages sitting behind the thousands of people who take IELTS every month, you've probably had this thought at 2am before your test: my accent is going to cap my score, no matter how good my English is.
It's one of the most common fears in IELTS prep, and it drives a lot of wasted energy — students spend months trying to sound American or British, imitating YouTubers, recording themselves obsessively hunting for "foreign" sounds. Some avoid speaking altogether because they're embarrassed by how they sound.
Here's the correction: that fear is based on a misreading of what the Pronunciation criterion actually measures. The official IELTS Speaking band descriptors do not ask you to sound native. They ask whether you're easy to understand. Those are not the same thing, and the gap between them is where most of your available marks are sitting, unclaimed.
Pronunciation is one of four criteria examiners score you on, alongside Fluency and Coherence, Lexical Resource, and Grammatical Range and Accuracy. Unlike the other three, it isn't about how much vocabulary or how many grammar structures you can produce. It's about control of specific features: word stress, sentence stress, chunking, intonation, individual sounds, and linking. Examiners are trained to listen for these features and judge how consistently you control them — not to judge whether your voice sounds like it belongs to someone from London or Sydney.
That distinction matters because it changes what you should actually be practising. Below is what changes at each band, straight from the descriptors.
At Band 5, a candidate uses some effective pronunciation features, but not consistently — causing real strain for the listener. At Band 6, that control becomes a mix of positive features, generally understood, with the listener needing to make some effort at times.
The single biggest lever between these two bands is word stress on multisyllabic words. Say "comfortable" with the stress on the wrong syllable and a listener has to work harder to recognise the word at all, even though every individual sound in it is correct. Wrong word stress breaks comprehension more severely than most individual sound errors, because English listeners rely on stress patterns to parse words in real time. This is one of the most underrated fixes in IELTS prep: most students never practise it deliberately because it isn't a "sound," it's a pattern.
The second lever at this stage is consistent production of the specific sounds that don't exist in your own first language. This is not a generic list of "hard English sounds" — it's different for every L1. A Mandarin speaker's interference sounds are not the same as a Spanish speaker's, which are not the same as an Arabic speaker's. Diagnosing your own specific list is far more useful than drilling a general pronunciation checklist aimed at nobody in particular.
Moving from 6 to 7, the descriptor shifts from "uneven control" to flexible use of features, more consistently, even though some accent effects remain.
Two skills dominate here. The first is sentence stress for meaning: stressing the content word that actually carries the point of the sentence, rather than stressing every word equally or defaulting to the last word out of habit. Compare "I really enjoyed the museum" said with stress on "enjoyed" (meaning I didn't just visit, I liked it) with the same sentence stressed on "museum" (meaning that place specifically, not the park). Same words, different meaning, purely through stress.
The second is chunking, or thought grouping — pausing at natural grammatical boundaries instead of pausing randomly or not pausing at all. This is what makes fluent speech easy to follow in real time. A student with perfect grammar who pauses in the wrong places still sounds hard to track.
At Band 8, a candidate uses a wide range of pronunciation features with only occasional lapses, is very easy to understand overall, and L1 accent has minimal effect on intelligibility.
The features that separate 7 from 8 are intonation for attitude and function, and linking. Intonation here means rising pitch for genuine uncertainty or questions and falling pitch for statements, used to convey actual meaning rather than a flat monotone or an artificial sing-song pattern some students adopt because they think it sounds "more native." It doesn't; it sounds performed. The goal is intonation that matches what you actually mean, not a memorised melody.
Linking is natural consonant-to-vowel connection across word boundaries. You don't need slang-style reductions; those aren't required and won't raise your score. What matters is that speech flows instead of sounding chopped into separate blocks.
This is the section that should end the "my accent will cap my score" fear for good. At Band 9, the descriptor states that a candidate uses a full range of pronunciation features with precision and subtlety, is effortless to understand throughout, with minimal L1 accent effect.
Minimal effect — not zero, not eliminated, not native-like. A Band 9 speaker can still sound recognisably non-native. Nowhere in the nine-band scale does the descriptor require you to erase your accent. It requires that your accent not get in the way of being understood. Those are fundamentally different goals, and confusing them is what sends thousands of students down the wrong training path every year.
Read the official IELTS Speaking band descriptors end to end and you will not find the phrase "native-like." You will not find "no accent." What you will find, repeated across the top bands, is language like "easy to understand" and "L1 accent has minimal effect on intelligibility." The examiner is assessing intelligibility and control of specific features, not origin.
Given all that, it's worth asking why so many students still spend their prep time on the wrong target. The usual pattern is obsessing over individual sounds — the "th" sound is the classic example — while ignoring word stress and sentence stress, which cost far more in the actual band score.
There's a reason this happens: individual sounds are easy to notice and easy to feel self-conscious about. Word stress errors are quieter — you don't feel them the same way, even though they do more damage to comprehension. A student who nails every "th" sound but stresses "comfortable" wrong, and pauses randomly instead of at clause boundaries, will still sound harder to follow than one who mispronounces a few consonants but has strong stress and chunking.
The problem with all of this is that most students are guessing. You know you have "an accent," you don't know which specific feature — word stress, sentence stress, chunking, intonation, or a specific L1-interference sound — is actually the one holding your band down.
This is exactly what Actually Speak's Accent Scan is built to answer. Instead of a generic accent score, it analyses your own spoken answers against the actual features examiners are trained to listen for — word stress, sentence stress, chunking, intonation, linking, and the specific sounds your own first language tends to interfere with — and shows you, with data, which of those is genuinely costing you marks right now.
If you've been avoiding speaking practice because you're convinced your accent is the ceiling on your score, run Accent Scan on Actually Speak before you assume anything. It's a five-minute way to replace a guess with a diagnosis, and in most cases the answer is a lot more fixable than "my accent" — usually it's one or two specific features, trainable in weeks, not a wholesale reinvention of how you sound.
For students who want that diagnosis paired with hands-on correction, Matthew Fabling's 1:1 coaching at smoothenglishcompany.com works through exactly these features — word stress, sentence stress, chunking, intonation — with direct feedback on your own speech rather than generic drills. Actually Speak's broader IELTS Speaking practice system builds on the same band descriptors covered here, so what you train matches what's actually being scored.
Will IELTS examiners understand me if I have a strong accent?
In most cases, yes — intelligibility and accent strength are not the same thing. Examiners are trained to assess whether specific features like word stress, sentence stress, and chunking are under control, and a strong accent with those features intact is usually still easy to follow. The bigger risk to being understood is inconsistent word stress or random pausing, not accent alone.
Do I need to sound British or American to get a high IELTS speaking score?
No. The band descriptors never mention sounding native, British, or American at any band, including Band 9. What they ask for is being easy to understand, with L1 accent having only a minimal effect on intelligibility — you can sound recognisably non-native and still score at the top of the scale.
What is word stress and why does it matter for IELTS?
Word stress is which syllable you emphasise in a multisyllabic word. It matters because English listeners use stress patterns to recognise words quickly, so getting it wrong causes more comprehension breakdown than most individual sound errors — and it's the single highest-leverage fix between Band 5 and Band 6.
Can I get Band 9 in Speaking with a noticeable accent?
Yes. The Band 9 descriptor explicitly states that L1 accent has minimal effect on intelligibility, not that it's absent. A Band 9 speaker can still sound clearly non-native; what's required is precise control of stress, intonation, chunking, and linking so that accent never interferes with being understood.
What is Accent Scan?
Accent Scan is a diagnostic tool on Actually Speak that analyses your spoken answers against the real features IELTS examiners score — word stress, sentence stress, chunking, intonation, linking, and the specific sounds your first language tends to interfere with — and shows you, with data, which ones are actually affecting your band, instead of leaving you to guess generically about "my accent."