IELTS Speaking Part 3: Talking About Numbers You Do Not Know

The examiner asks whether people in your country read less than they did twenty years ago. You have no figures. You have an impression, and an impression feels like it is not enough, so the answer stalls somewhere around the third word.

The stall comes from a false premise. Part 3 is not a general knowledge quiz; it is a discussion, and what is being assessed is how you compare, speculate and explain in English. Candidates who accept that stop searching for a statistic and start answering the question.

This guide gives you the structure for data questions, the phrases that keep approximation honest, and two full sample answers built on that structure. The questions and answers here are original practice material, not official or recalled questions, and no answer has been assessed or given a band.

Why Part 3 numbers questions test language, not data

Part 3 follows the Part 2 topic into a more general and abstract discussion and lasts four to five minutes. Speaking is scored on four criteria that carry equal weight: fluency and coherence, lexical resource, grammatical range and accuracy, and pronunciation. None of the four is a fact-check.

The descriptors show what is actually rewarded. Band 7 lexical resource includes Effective use of paraphrase as required, and band 7 fluency and coherence includes Flexible use of spoken discourse markers, connectives and cohesive features. Those two lines describe a candidate who can reframe a question, mark a comparison, and keep going, which is exactly what an unanswerable-looking data question invites.

This is not permission to invent things. A confident wrong figure invites a follow-up you cannot survive, and it sounds thin, because there is nothing behind it. The honest structure below is also the easier one to sustain for thirty seconds.

Think in three sentences: scope, direction, reason

Every data answer can be built from three sentences, and the order matters: limit the claim first, so the examiner knows what kind of answer is coming.

Sentence	Job	Weak version	Strong version
One	Limit the claim	I don't know.	I don't have figures for that, but I can tell you what I see around me.
Two	Give the direction, not the size	It's less.	It's clearly down, though probably not as much as people assume.
Three	Give one cause	Because of phones.	Mostly because a phone is what people reach for when they have five minutes.

That is roughly twenty-five to thirty-five seconds of speech, and it answers the question without a single number. If the examiner wants more, the third sentence hands them a thread to pull, which is the outcome you want: a follow-up you can answer is worth more than a monologue that ends in silence.

Approximation phrases that stay honest

The phrases below are not filler. Each one tells the listener exactly how much weight your claim can carry, which is the difference between approximation and invention.

Phrase	What it commits you to	Example
in my experience	your own small sample, nothing more	In my experience, most people I know read on a phone now.
a small share	clearly less than half, no figure	Only a small share of the people I know still buy a newspaper.
roughly	an estimate you would defend if asked	The number of readers is roughly the same; the format changed.
somewhere between	a range you can justify	For people my age it's somewhere between a third and a half of that, though that is a guess from my own circle.
the general direction	up or down only	I'm confident about the general direction, not the size.

Two rules keep these safe. Hedge the size, never the direction: It has definitely gone down, though I couldn't tell you by how much is stronger than a vague maybe it's a bit less. And never borrow authority: if you cannot name the study, do not say according to a study, because that sentence is the one that turns an honest answer into a false one.

IELTS Speaking Part 3 sample answers you can rebuild

Two original questions, answered with the three-sentence pattern. The annotation after each one shows where the pattern sits.

Question one: Do you think people read less than they did twenty years ago? I don't have any figures for that, so I can only go by what I see. In my experience it's clearly down, but mainly in one specific way. People still read a lot; they just read on a phone, so the amount has stayed roughly the same while the format changed, mostly because a phone is what you reach for when you have five minutes.

Question two: Is it more expensive to be young today than it was for your parents? That's hard to compare directly, because what people spend money on has changed. From what I can see, housing has become much harder and almost everything else has become cheaper, so the answer depends on whether you need somewhere to live. I think that's why the argument never settles: two people can look at the same decade and describe completely different lives.

Question one limits itself to personal observation, then gives a direction, then explains the direction. Question two does something slightly harder: it accepts that the question has no single answer and says what the answer depends on, which is a legitimate move in a discussion and reads as control rather than evasion.

Trend verbs and the adverbs that keep them accurate

Once you have a direction, the verb decides how precise you sound, and the adverb decides whether the precision is honest.

Verb	What it says	Adverb that keeps it honest
rise, climb	movement upwards	steadily for gradual change, sharply only if you can name the cause
plateau, level off	stopped rising, still high	after a long climb
dip	a small, temporary fall	slightly, or briefly
stay flat	no real change	more or less, which admits your view is approximate
fall back	returns to an earlier level	to where it was
overtake	one line passes another	in the last few years

Tense carries meaning too. It has risen steadily since the nineties covers a period that runs up to now; It rose steadily through the nineties covers a period that finished. Noun forms are worth having ready, because they let you vary the sentence: a steady rise, a slight dip, a sharp fall. Pairing a strong verb with a weak adverb, or the reverse, is the most common way an otherwise good answer starts to sound like a news report nobody wrote.

Comparing two groups without a single figure

Comparison questions are the easiest place to fall into a fake statistic, because the sentence shape invites one. Five frames avoid it.

• Young people spend more on rent than their parents did at the same age.
• Older people still read newspapers, whereas younger people mostly read on phones.
• The opposite is true for people who grew up without the internet.
• By contrast, the cost of almost everything else has fallen.
• The gap seems to be widening, though I couldn't tell you by how much.

One grammar point is worth the practice: compared with takes a noun, so compared with their parents is right, while than takes a clause, so than their parents did is right. Mixing them, as in compared with their parents did, is the kind of error that is easy to make while thinking about the argument rather than the sentence.

An eight-question drill and a self-check for invented statistics

Answer each question in three sentences, out loud, with a timer. Aim for thirty seconds on the first pass and forty-five on the second.

1. Do you think people in your country read less than they did twenty years ago?
2. Are young people today better at saving money than their parents were?
3. Has the way people shop changed much in your lifetime?
4. Do you think cities are getting more crowded?
5. Is it cheaper to eat at home than to eat out where you live?
6. Do people in your country take more holidays than they used to?
7. Which costs more, owning a car or using public transport?
8. Do you think the number of people learning a second language is growing?

Then run the self-check on your recording. Any figure you cannot source gets replaced with a direction. Any percentage gets deleted and re-said as more, less or about the same. Any according to gets a name or gets dropped. Any sentence beginning Statistics show gets cut entirely, because it is the sentence examiners hear most often and believe least. If a clip survives all four checks, you have an answer that sounds informed without containing a single unsupported number.

Frequently asked questions

Do I need to give real statistics in IELTS Speaking Part 3?

No. The four scoring criteria describe language, and a direction plus a cause answers the question without a figure. Inventing a percentage is riskier than admitting the limit of your knowledge, because it creates a claim you have to defend in the next question.

How long should a Part 3 answer be?

Three to five sentences, about thirty to forty-five seconds, is usually enough. Part 3 gives four to five minutes to the whole discussion, so one long answer spends time that later questions need. Answer, then stop and let the examiner ask the next one.