IELTS Speaking Part 1: a short answer is not a safe answer

Part 1 of the IELTS speaking test asks you about your hometown, your job, your routine, whether you like cooking. The questions are deliberately easy, and that is exactly what makes them dangerous.

An easy question invites an honest, accurate, complete answer:

Do you enjoy cooking? Not really, no. I usually order in.

Nothing about that is wrong. It is true, it is grammatical, it is clear. It is also eight words, and eight words is not enough language for anyone to place you in a band.

This post is about that specific failure — under-production — because it is the one that costs Part 1 marks, and it is almost never described in terms of what the examiner is actually short of. We build the engine that scores speaking on Hilingo, so this is written from the marking side. For the descriptors themselves, the source is ielts.org.

Part 2 and Part 3 fail in different ways, and they have their own posts: running dry in the long turn and developing an abstract argument.

One band per criterion, for the whole interview

Here is the structural fact that everything else follows from.

IELTS speaking is marked on four criteria, weighted equally: fluency and coherence, lexical resource, grammatical range and accuracy, and pronunciation. You do not get a band for Part 1 and another for Part 2. You get one band per criterion for the entire fourteen minutes, and your speaking band is the average of those four.

So Part 1 is not a round you can win. There is no Part 1 score. What Part 1 does is contribute — or fail to contribute — to the pool of language the examiner uses to place you on all four criteria.

That reframes the goal. In Part 1 you are not trying to give good answers. You are trying to produce enough language to be banded on.

Why six words cannot be marked

Look at what each criterion actually requires, and it becomes obvious that two of the four are structurally incapable of being shown in a one-sentence answer.

Criterion	What "Not really, no. I usually order in." shows	What it needs
Fluency and coherence	One clause. Nothing had to connect to anything.	Speech that keeps going and holds together across several ideas
Lexical resource	Two content words, both high-frequency	Range, and the ability to paraphrase around a topic
Grammatical range and accuracy	Present simple, once	More than one structure, used accurately
Pronunciation	A genuine sample — this one does work	Intelligible sound across connected speech

Fluency and coherence is the clearest case. Coherence is a property of ideas that follow from each other. A single sentence has nothing to follow from, so it cannot demonstrate coherence at all — not badly, not at all. You have not scored low; you have produced nothing to score.

Grammatical range is the same. Range means a mix of structures. One present-simple clause is evidence of exactly one structure, and no examiner can infer the rest.

We ran into this from the other side when building the scorer. Marking grammatical range answer-by-answer punished every candidate for the shape of Part 1, because a true Part 1 answer — "Yes, I do. Mostly at weekends." — is too short to show range no matter how good the speaker is. The engine now pools the whole interview's transcript and scores that criterion once, which is what a human examiner does and cannot do otherwise. The consequence for you is the point of this post: a thin Part 1 does not get marked leniently, it gets absorbed into a pool it barely contributed to.

The present-tense trap

There is a reason under-production in Part 1 specifically flattens your grammar, and it is built into the questions.

Part 1 asks about habits and preferences. Do you like...? Do you often...? What kind of...? Where do you live? Answer any of those naturally and accurately, and you will produce the present simple. Answer twelve of them naturally and accurately, and you will produce twelve present simples.

You have been fluent, correct and completely honest, and you have shown the examiner one tense.

This is the most fixable thing in Part 1, and the fix is not complexity. It is a second time frame:

Do you enjoy cooking? Not really, to be honest — I usually end up ordering in after work. I used to cook a lot more when I lived with my parents, mainly because there was someone to cook for. If I had more time I think I'd get back into it.

That is three time frames — present habit, past habit, conditional — in three sentences. No rare vocabulary, no clause you might abandon halfway. The grammatical range is doing the work, and it came from adding when and if, not from adding difficulty.

How much is enough

Not a word count — a content count. Each Part 1 answer should carry:

1. The direct answer. Say yes or no or "it depends" first. The examiner should never have to wait to find out what you think.
2. A reason. This is where connected speech starts, and therefore where fluency and coherence become markable.
3. One shift. A past example, a comparison, a condition, an exception. This is your grammatical range and it is the part candidates skip.

Three or four sentences, roughly. If you want a check that works while you are speaking: stop when you have said something in a tense other than the present.

The other ceiling: do not speak for ninety seconds

Under-production is the dominant failure, but the opposite one is real and worth naming, because advice like the above can push people into it.

Part 1 is short — a few familiar topics over four to five minutes — and the examiner has a set of questions to get through. A ninety-second answer to "Do you like tea?" does two things to you. It reads as prepared rather than spontaneous, and it invites the examiner to move you on, which means an answer of yours ends mid-idea.

Part 1 is warm-up-shaped for a reason. Extend every answer; do not turn any of them into a speech. The long turn is Part 2, and it has its own rules.

Memorised answers are most obvious here

Part 1 topics are predictable. Hometown, work, study, food, weather, free time — they recur, everyone knows they recur, and so a lot of candidates arrive with paragraphs learned.

It does not survive contact with the test, for three reasons.

The rhythm changes, audibly. Recited speech has different pausing, different stress and a different pace from spontaneous speech, and the change happens at the exact moment you start the rehearsed section. An examiner hears fourteen minutes of you; the join is not subtle.

It is self-defeating on the criterion it is meant to help. A memorised answer usually exists to deploy impressive vocabulary. But lexical resource is banded across the whole interview, so a polished sentence in answer three sits next to your real range in answer four — and the contrast reads as borrowed, not as range.

It breaks on the follow-up. Part 1 questions branch. You prepared "my hometown"; the examiner asks whether it has changed since you were a child. The prepared paragraph does not fit, and you now have to improvise from a standing start, having rehearsed the opposite skill.

Prepare ideas, not sentences. For each common topic, know two reasons and one concrete example you could reach for. That survives any phrasing of the question, and it produces the "one shift" above without any rehearsal.

What to do if you did not understand

Ask. "Sorry, do you mean...?" is not a penalty. It is spontaneous, appropriate language, and it is better on every criterion than an answer to the wrong question.

In Part 1 the examiner can repeat a question. What no one can rescue is a confident answer about something you were not asked, which reads as a comprehension failure and produces irrelevant language on top of it.

The check you can run on yourself

You cannot judge your own production while producing it — the same reason the IELTS mistakes you cannot see yourself are invisible. But you can measure it afterwards.

Record yourself answering ten Part 1 questions. Then, for each answer:

• Count the words. Under fifteen, consistently, means under-production and nothing else.
• Count the distinct tenses across all ten. If it is one, you have found your grammatical range ceiling, and it is not a knowledge problem — you know the past tense — it is a production habit.
• Count the restarts. Restarts and abandoned clauses are what fluency actually measures. Speed is not.

Three numbers, ten minutes. They will tell you more than another list of sample answers will.

See what your Part 1 answers are actually producing

The thing you cannot do for yourself is hear your Part 1 answers the way an examiner hears them — as a contribution to a whole-interview judgement rather than as individual replies that each sounded fine.

Hilingo gives you one full scored mock free, with no card. Speaking is assessed across Parts 1 to 3 with a band on each of the four criteria, and each answer comes back with its own transcript so you can count words, tenses and restarts on real timed answers instead of on a rehearsal. The report also flags which criterion is the least certain, rather than presenting four numbers as though they were equally solid.

Take a free scored mock

• IELTS Speaking Part 2 — the long turn, and what to do when you run dry
• IELTS Speaking Part 3 — developing and justifying an abstract answer
• The IELTS mistakes you cannot see yourself — why one criterion is usually a band below the rest
• The last week before your IELTS — what moves your band late
• IELTS band score chart — band to CLB per skill

Frequently asked questions

How long should IELTS Speaking Part 1 answers be?

Around three or four sentences — long enough to contain a direct answer, a reason and one shift into another time frame. One-sentence answers are the common failure: they are accurate but give the examiner almost no language to assess, and two of the four criteria cannot be demonstrated in a single clause at all.

Does IELTS Speaking Part 1 affect your band score?

Yes, but not as a separate mark. There is no Part 1 score. You receive one band for each of the four criteria across the whole interview, so Part 1 contributes language to that judgement. It cannot win you a band on its own, which is why under-producing there is expensive and over-performing there is not rewarded.

Is it OK to memorise answers for IELTS Speaking Part 1?

It is the worst part of the test to do it in. The topics are predictable, so examiners hear rehearsed answers constantly, and the rhythm of recited speech differs audibly from spontaneous speech. It also breaks on follow-up questions. Prepare two reasons and one example per common topic instead of full sentences.

What if I don't understand a question in IELTS Speaking Part 1?

Ask for clarification. Doing so is normal, spontaneous language and is not penalised. Answering the question you assumed was asked is worse on every criterion, because it produces irrelevant content as well as reading as a comprehension failure.

Should I use difficult words in IELTS Speaking Part 1?

No. Lexical resource rewards range used accurately, and an unfamiliar word used slightly wrong damages the criterion more than a plain word used correctly. In Part 1 the more reliable lever is grammatical range — adding a past example or a conditional — because it is easier to control under pressure.

What topics come up in IELTS Speaking Part 1?

Familiar personal ones: where you live, your work or studies, your routine, food, weather, free time, transport. The predictability is why memorisation is tempting and why it is so visible. Knowing the topics is useful for preparing ideas, not for preparing scripts.