PTE Retell Lecture: How to Select Numbers and Examples

Numbers feel safe in PTE Academic Retell Lecture because they are concrete. Yet an answer packed with dates, percentages, and sample sizes can miss the lecture's argument. A number earns space only when it establishes scale, comparison, change, or a central finding.

Pearson's current format allows a lecture of up to 90 seconds, followed by 10 seconds to prepare and 40 seconds to speak. The task is to retell the lecture in your own words. You need a selection test: does this detail carry the main point, support it, or merely illustrate it?

Practice lecture: numbers with different jobs

This is an original practice text:

Many museums assume that longer opening hours automatically increase access, but timing must match visitor availability. A city museum tested late closing on Thursdays for twelve weeks. Its normal Thursday attendance was about 420 visitors. During the trial, average attendance reached 560, an increase of roughly one third. However, most additional visitors arrived between 6 and 7 p.m.; the final two hours attracted fewer than 40 people each. Interviews with 120 visitors suggested that the early-evening window helped workers visit after office hours. For comparison, a second museum extended Sunday hours and recorded only a 4 percent increase. These figures indicate that extension alone is not enough: the placement of the extra hours matters. The trial was short and involved only two museums, so it does not establish a rule for every institution.

The lecture contains many numbers, but they have different roles:

Main finding versus memorable detail

The central finding is not “560 people attended.” It is that when extra hours are placed may matter more than adding hours generally. Attendance growth and its concentration in the early evening support that point.

Use verbs that preserve the relationship:

Attendance rose from about 420 to 560, with most of the additional visits concentrated between 6 and 7 p.m.

The pattern supports the lecturer's claim that timing matters.

The first sentence reports; the second interprets the role the lecturer gives the numbers. Neither says the experiment proves a universal causal law.

Preserve units and bases

Write units in abbreviated notes:

420→560 visitors/Thu (~+1/3)

late final hrs <40 each

Museum2 Sun ext +4%

Without visitors per Thursday, 420 and 560 could be totals across the entire trial. Without each, fewer than 40 could refer to the two hours combined. Without Sunday and the second museum, the 4 percent figure can be attached to the wrong comparison.

When speaking, restore the unit naturally. Do not invent precision: about 420 should remain approximate, and roughly one third should not become exactly 33.3 percent.

Raw numbers or interpreted comparison?

You rarely need both.

Attendance increased from about 420 to 560.

Attendance increased by roughly one third.

Either can efficiently establish scale. Giving both may help if delivered cleanly, but it uses time. Choose the form that best serves the central point and that you recorded reliably.

Do not confuse an absolute difference with a relative one. The increase is about 140 visitors, or roughly one third of the original 420. It is not 140 percent.

What an example does

Examples can illustrate, support, contrast with, or qualify a claim.

The second museum is a comparison case:

By contrast, a Sunday extension at another museum produced only a 4-percent increase.

That contrast supports the idea that placement matters. It does not establish that Thursday evenings are always superior to Sundays; the museums and contexts may differ.

The visitor interviews offer a possible explanation:

Interviews suggested that the early-evening period suited people leaving work.

Use suggested, not proved. Self-reported explanations and a small trial do not license a universal cause.

Numerical verbs and boundaries

Rise from X to Y states starting and ending values.

Rise by X states the size of the increase.

Reach X states the resulting level without naming the starting point.

Account for describes a share of a whole.

Concentrate in identifies where observations gather.

Exceed compares levels but does not itself describe change.

Compare:

Attendance rose from 420 to 560.

Attendance rose by about 140.

Attendance reached 560.

Switching to and by is a small preposition error with a large numerical effect.

A 40-second model

The lecture questions whether simply extending museum hours improves access. In a twelve-week Thursday trial, average attendance rose from about 420 to 560, but most extra visitors came between 6 and 7 p.m., while the final hours attracted few people. A second museum's Sunday extension produced only a 4-percent increase. The speaker therefore argues that the timing of extra hours matters, not just their length. However, the evidence covers only two museums and should not be treated as universal.

This response keeps four numerical ideas because each performs a different job: duration, central change, time concentration, and comparison. It drops the interview sample size because that number does not change the argument.

A deletion test for notes

For each number, ask:

1. If I delete it, does the main conclusion lose its basis?
2. Does it show the scale or direction of the main result?
3. Does it create the key contrast or caveat?
4. Can I state its unit and comparison base confidently?

If all answers are no, omit it. If the fourth answer is no, paraphrase safely or omit it rather than attaching the figure to the wrong measure.

Common mistakes

• Listing all figures chronologically. Reorganize them around the claim.
• Dropping units. A number without its measure may become misleading.
• Changing from to by. Preserve the arithmetic relationship.
• Treating an example as proof. Report what it illustrates and keep scope limits.
• Correcting a number from outside knowledge. Retell the lecture, not what you think the figure should be.
• Inventing exactness. Keep about, roughly, nearly, and fewer than.
• Sacrificing the conclusion for details. One useful number plus the main point is better than six disconnected figures.

Worked triage with conflicting numerical attractions

Consider this original mini-talk:

A university tested shorter orientation videos in four departments. The original version lasted 28 minutes, while the revised version lasted nine. Completion rose from 54 to 76 percent. The largest gain occurred among part-time students, although the talk did not give that subgroup's exact size. More viewers completed the revised quiz as well, but the change was only two percentage points. The speaker concludes that reducing length may improve completion, while warning that the departments chose voluntarily whether to participate.

The strongest notes are 28→9 min, completion 54→76%, part-time biggest gain/no n, and voluntary departments = selection caveat. The two-point quiz change is accurate but secondary because the conclusion concerns completion.

A controlled retell is:

The speaker describes a four-department trial in which shortening an orientation video from 28 to nine minutes coincided with a rise in completion from 54 to 76 percent, especially among part-time students; however, the subgroup size was not given and voluntary departmental participation limits a causal conclusion.

Why say coincided with instead of caused? Voluntary participation leaves a selection difference, and the speaker uses may. Why retain both time values? They establish the scale of the intervention. Why omit the quiz result? It does not materially develop the stated conclusion.

Answer check

1. The completion gap is 22 percentage points, not 22 percent.
2. The video became 19 minutes shorter; it did not fall by 19 percent.
3. The part-time result should remain qualitative because the exact subgroup figure was not provided.
4. “Across all universities” is unsupported because the talk covers four self-selecting departments.

Each answer follows from a named value or limit. If you cannot restore a unit, population, or comparison base from your notes, use a faithful qualitative relationship or leave the number out.

Quick practice

Your notes say: program 8 schools / reading +12% / strongest age 9–10 / one term / no control group.

A safe retell is:

The lecturer reports a 12-percent reading improvement across eight schools, especially among pupils aged nine to ten, but cautions that the one-term study lacked a control group.

Do not say the program caused the improvement unless the lecture supports that causal conclusion. Do not change 12 percent to 12 percentage points without knowing the measure.

Create a new 35-second response using only three of the five notes. Explain why each retained detail contributes to finding, scope, or limitation.

For argument links, see PTE Retell Lecture: connect claims and evidence. For note structure, use PTE Retell Lecture signposts. The PTE Academic Vocabulary hub and Exam English hub contain the larger pathway.

Sources

• PTE Academic Speaking & Writing test format — Pearson
• PTE Academic speaking test tips — Pearson

Lyra Practice is not affiliated with or endorsed by Pearson.