Explain why a technique supports a scoring trait, apply its limits, and deliver complete responses to different images.
Learning path
- Task and scoring
- Content techniques and limits
- Delivery techniques
- Preparation and transfer
- Complete demonstrations
- Review
Examples are teaching illustrations, not officially scored candidate responses.
Task and scoring
Describe Image: building a stronger answer
- Recognise the scoring differences
- Build and explain the relationships
- Deliver the answer clearly
Explanation
In Describe Image, a correct number is only the beginning. A response can name every number and still leave the relationships unexplained. This lesson shows how to turn those isolated facts into an organised description, and how to deliver that description clearly. We will use the scoring criteria to diagnose a weak answer, improve it step by step, test the method on different images, and finish with three complete spoken models.
You have twenty-five seconds to prepare and forty seconds to speak. Wait for the short tone, then begin when the microphone opens; you can record only once. That time limit makes planning part of the method.
Three traits, each scored from 0 to 5
| Criterion / example | Meaning / explanation |
|---|---|
| Content · 0–5 | What the image shows |
| Pronunciation · 0–5 | How clearly words are heard |
| Oral fluency · 0–5 | How speech holds together |
Explanation
The official traits are content, pronunciation and oral fluency. Communication is the overall purpose; it is not a fourth separately scored trait here. Content concerns what your answer communicates about this image. Pronunciation concerns how clearly the words can be understood. Oral fluency concerns the continuity and phrasing of your speech. These distinctions matter. A polished delivery cannot supply a relationship you never explained.
Equally, a strong written paragraph cannot demonstrate how you actually pronounce it. We will improve the description first, then work on the delivery.
Content: what separates 1, 3 and 5
| Band | Observable qualities | What to listen / look for |
|---|---|---|
| 1 | Basic features | Connections left unclear |
| 3 | Most key features | Conclusion or implication |
| 5 | All elements and relationships | Development + conclusion / implications |
Plain-language summary · Pearson PTE Core Score Guide, pp. 27–28
Explanation
At content band one, basic features appear, but their connections or implications remain unclear. Band three covers most key features and addresses a conclusion or implication. Band five describes all elements and their relationships, addresses possible development, and gives a conclusion or implications. Development must still follow the evidence: it does not mean inventing a future trend or a hidden cause. The practical difference is coverage plus development.
On a simple four-category chart, I can cover all four categories and explain how they relate. On a process, I would explain the stages and their sequence. A memorised sentence about percentages would not serve that image. The method must follow what is actually shown. At zero, the response mentions disjointed elements or fails to address the image properly, for example through unrelated memorised material.
Content zero means no score points for that response, so polished filler cannot replace relevant description.
Content techniques and limits
A list of correct facts can still be weak
Main commuting mode
| Category | Value |
|---|---|
| Cars | 40 |
| Buses | 30 |
| Trains | 20 |
| Bicycles | 10 |

Respondents (%) · One main mode per person
Explanation
Here is our example: the percentage using each main commuting mode. Hear a list of accurate facts.
Response model
Cars are forty percent. Buses are thirty percent. Trains are twenty percent. Bicycles are ten percent.
The figures match. But the answer does not identify what is being measured or develop a relationship. The lower content bands describe this weakness. We are diagnosing features, not awarding this recording an official band. The first technique connects the facts while retaining coverage.
Technique 1: choose a relationship to explain
Main commuting mode
| Category | Value |
|---|---|
| Cars | 40 |
| Buses | 30 |
| Trains | 20 |
| Bicycles | 10 |

Respondents (%) · 40 ÷ 10 = 4 times the bicycle share
Explanation
The first technique is to turn separate figures into a supported comparison. This targets content, because the rubric assesses relationships as well as individual details. Here, I select the largest and smallest categories. Cars account for forty percent, compared with ten percent for bicycles. Cars therefore have four times the share of bicycles. The original answer named two values; this version explains their relationship. That is the improvement.
I would still cover buses and trains, because one comparison does not replace coverage. Use a ratio only when the values share the same measure and the calculation is straightforward. Otherwise, a simple higher-than comparison works.
A turning point changes the comparison
Annual library visits
| Category | Value |
|---|---|
| 2021 | 20 |
| 2022 | 40 |
| 2023 | 30 |
| 2024 | 50 |

Visits (thousands) · Overall growth, with a fall in 2023
Explanation
Let us apply technique one to a line graph. Library visits rise from twenty thousand to forty thousand, then fall to thirty thousand before recovering to fifty thousand. Saying visits increased steadily would erase the fall. A stronger comparison is: visits increased overall, despite a decline in the middle of the period. The endpoints establish the overall direction; the turning point explains the pattern between them.
This supports content by preserving a significant relationship. Naming only the highest and lowest values would leave that relationship underdeveloped.
A flat period is also information
Annual library membership
| Category | Value |
|---|---|
| 2021 | 20 |
| 2022 | 30 |
| 2023 | 30 |
| 2024 | 40 |

Members (thousands) · Growth pauses between two increases
Explanation
Now change the shape of the line. Membership rises from twenty to thirty thousand, stays at thirty thousand for another year, then reaches forty thousand. There is no decline here. The useful relationship is a pause in growth. I would say: membership rose overall, with a year of no change between the two increases. Reusing the previous phrase, despite a decline, would introduce something this graph does not show.
The technique transfers because we inspect the relationship again; the sentence itself is not a template to copy unchanged.
A higher percentage can describe fewer people
| Branch | Participation rate / base | Participants |
|---|---|---|
| A | 60% of 100 members | 60 |
| B | 40% of 200 members | 80 |
Compare rates separately from headcounts; the group sizes differ.
Explanation
One more comparison needs a different check: what is the percentage out of? At branch A, sixty percent of one hundred members attend a workshop. At branch B, forty percent of two hundred attend. Branch A has the higher participation rate, but branch B has more participants: eighty compared with sixty. Saying A has more people because its percentage is higher would misstate the content. If the group sizes are unavailable, compare the rates only.
A ratio of percentages cannot establish a ratio of headcounts when the underlying groups differ.
Technique 2: turn related details into an overview
Main commuting mode
| Category | Value |
|---|---|
| Cars | 40 |
| Buses | 30 |
| Trains | 20 |
| Bicycles | 10 |

Respondents (%) · Bus 30 + train 20 = 50; no overlap
Explanation
The second technique is to group related details into one meaningful comparison. It supports content development while letting you cover several elements within the speaking time. This chart records one main mode per respondent, so bus and train form separate groups. Their thirty and twenty percent combine to fifty percent for public transport. That is ten percentage points above cars.
This changes our overview: cars lead individually, but public transport leads when its two categories are combined. We have covered both categories and explained their combined significance. Adding percentages is optional. If categories overlap, adding them can create an inaccurate content claim. In that case, compare them separately.
Multiple selections make addition unsafe
Travel modes used during the week
| Category | Value |
|---|---|
| Bus | 60 |
| Train | 50 |

Respondents (%) · Multiple selections; overlap unknown
Explanation
Technique two has an important limit. This survey asks which travel modes people used during the week, and respondents may select several. Sixty percent selected bus and fifty percent selected train. Adding them gives one hundred and ten percent, but that is not the percentage who used public transport. Some people appear in both groups. A supported comparison is: a higher percentage of respondents used buses than trains.
We cannot calculate the combined unique share without knowing the overlap. Grouping helps content only when the combined claim remains accurate.
A shared total makes grouping meaningful
One monthly budget
| Category | Value |
|---|---|
| Rent | 40 |
| Food | 30 |
| Transport | 20 |
| Savings | 10 |

Budget (%) · Rent + food = 70; other categories = 30
Explanation
Here is a case where grouping does work. A monthly budget allocates forty percent to rent, thirty to food, twenty to transport and ten to savings. These are separate portions of the same total. Rent and food together account for seventy percent, leaving thirty percent for transport and savings. That comparison connects all four parts and develops the overview. We should not conclude that the household is struggling financially, because income adequacy is not measured.
The grouping technique supports content through an accurate relationship, not through a dramatic story.
Technique 3: support the conclusion with evidence
Unsupported cause
Cars are cheaper, so more people choose them.
Evidence-based conclusion
Public transport has a larger combined share: 50% versus 40% for cars.
Explanation
The third technique is to build your conclusion from the relationship you have demonstrated. This addresses content development: the ending tells the listener what the details show together. Cars are cheaper, so more people choose them, introduces a cause the chart never measured. Instead, I can conclude that public transport has a larger combined share than cars, despite cars being the largest individual category. The two figures support that conclusion directly.
It adds meaning without inventing information. With a different image, the appropriate development might be a visible change over time or the result of a process. Choose the ending from that image, not from a memorised prediction.
A map supports location, not motive
| Position | Feature |
|---|---|
| Centre | Station |
| North | Library |
| East | Park |
| South | School |

Schematic location map · North is up · Positions support an overview, not a motive.
Explanation
Technique three also applies to a map. The library is north of the station. A park is east of the station, and a school is south of it. Together, these locations place the station between the library and school, with the park to its east. That is a supported spatial overview. Saying the station was built to serve schoolchildren introduces a motive the map never gives. For content, describe the arrangement and relevant connections you can see.
We do not need to invent an explanation for why the town was designed this way.
A process may contain an alternative route
Starting sequence: Collection<br>then sorting
Decision: Address valid?
| Condition | Next step |
|---|---|
| Yes | Delivery |
| No | Return to sender |

The address check determines two possible outcomes.
Explanation
A process needs a different kind of development. This parcel diagram shows collection followed by sorting. Parcels with a valid address go to delivery; parcels with an invalid address return to the sender. Describing only collection, sorting and delivery would miss the branch. A stronger explanation states the condition that determines the route. That develops content because it explains how the parts connect.
It would also be inaccurate to conclude that every parcel reaches its recipient. Here the outcome depends on the address check, so the conclusion must preserve that condition.
Word choice: accuracy, not a magic phrase
| Evidence | Accurate claim | Why |
|---|---|---|
| 40% and highest category | Largest individual share | Highest does not imply over half |
| 60% of respondents | Most respondents / a majority | The share exceeds half |
No published fixed point bonus for “largest” or penalty for “most”.
Explanation
This also explains the issue with most and largest. In this chart, the largest single category is cars. Calling that share a majority would be inaccurate, because forty percent is below half. The problem is the claim, not a forbidden word. If cars represented sixty percent in another chart, most respondents would be accurate. Pearson publishes no rule saying that largest earns a point or most loses one.
The public criteria describe the quality of the response; they do not disclose a word-by-word deduction formula. Use wording that preserves the meaning of the evidence.
An increase has two different numerical descriptions
| Claim | Calculation | Meaning |
|---|---|---|
| 20% → 30% | State both endpoints | Accurate without arithmetic |
| +10 percentage points | 30 − 20 = 10 | Difference between shares |
| +50% relative increase | (30 − 20) ÷ 20 = 0.5 | Increase relative to the initial share |
Teaching example: online share of all bookings. “Up by 10%” describes a different change.
Explanation
There is another wording trap that changes the meaning. Online bookings rise from twenty percent to thirty percent of all bookings. The difference is ten percentage points. Relative to the original twenty percent, the increase is fifty percent. Both statements are accurate, but they describe different calculations. Saying increased by ten percent would imply a different end value. If that calculation distracts you, say the share rose from twenty to thirty percent.
Content benefits from the accurate change; complicated arithmetic is not a scoring requirement.
The revised content has a clear job for every sentence
| Criterion / example | Meaning / explanation |
|---|---|
| Identify | Topic + measure |
| Compare | Cars 40% versus bicycles 10% |
| Connect | Bus 30% + train 20% = 50% |
| Conclude | Public transport 50% exceeds cars 40% |
Every sentence has a content job; these notes are not the spoken script.
Explanation
Returning to the commuting chart, we can now see the description taking shape. These notes show the information each part must carry, rather than a compulsory sentence order. The opening identifies what is measured. Cars and bicycles establish the largest and smallest individual shares. Buses and trains supply the combined public transport comparison. The conclusion explains why that combined share changes the overview.
In the complete model, I will introduce cars, connect buses with trains, add bicycles, and then conclude. Each sentence adds information the listener needs. This is stronger than the original list because it preserves the figures and explains their relationships. It illustrates features of stronger content, but a predicted five out of five would require an actual assessment of the whole response.
Delivery techniques
Oral fluency: what separates 1, 3 and 5
| Band | Observable qualities | What to listen / look for |
|---|---|---|
| 1 | Broken rhythm | Repeated disruptions |
| 3 | Mostly continuous phrases | May be uneven; no long pauses |
| 5 | Smooth phrasing | No hesitations or restarts |
Plain-language summary · Pearson PTE Core Score Guide, p. 40
Explanation
Content is now planned. The next issue is whether we can deliver it without repeatedly stopping to construct sentences. At fluency band one, broken rhythm and repeated disruptions make speech uneven. At band three, most words form continuous phrases, although the rhythm may be uneven; long pauses and staccato delivery are absent. At band five, phrasing is smooth, without hesitations, repetitions, false starts or disruptive simplification of speech sounds.
This does not ban normal connected speech: appropriate linking and sound changes also appear in the pronunciation criteria. The goal is connected speech with natural phrase boundaries. Racing through a sentence can make the next one harder to produce. We need a method for reducing the search for words while we speak. That brings us to the fourth technique: prepare the answer in short meaning groups.
Technique 4: plan and speak in meaning groups
Search and restart
Cars … cars account … for … forty percent of respondents.
One manageable group
Cars account for forty percent of respondents.
Explanation
Technique four is to plan a manageable meaning group. This targets oral fluency by reducing search and restarts. Hear a deliberately interrupted version.
Response model
Cars
Response model
cars account
Response model
for
Response model
forty percent of respondents.
Now the same intended wording, with the repeated opening removed.
Response model
Cars account for forty percent of respondents.
Category and value stay together. That is the fluency improvement. Now extend the thought with a planned comparison.
Response model
Cars account for forty percent of respondents, compared with ten percent for bicycles.
The comparison adds Content; the stable delivery supports Fluency. Keep these two improvements separate. A single interruption does not determine the whole fluency band.
Meaning groups adapt to a conditional process
After sorting, parcels with a valid address go to delivery.
| Phrase | Role |
|---|---|
| After sorting | Sequence |
| parcels with a valid address | Condition |
| go to delivery | Outcome |
The group follows the condition–outcome relationship, not a fixed word count.
Explanation
Apply the same grouping technique to the process. First hear a fragmented condition.
Response model
Parcels with
Response model
a
Response model
valid
Response model
address go to delivery.
Now hear the condition delivered as one phrase. I will also add the sequence and the other route, which develop Content.
Response model
After sorting, parcels with a valid address go to delivery. Parcels with an invalid address return to the sender.
Keeping the condition together supports Fluency. Adding the sequence and the second outcome develops Content; those are two separate changes. The group follows a meaningful relationship, not a fixed number of words or a ban on natural pauses. Keep both outcomes, because simplifying delivery must not remove an essential branch.
Pronunciation: what separates 1, 3 and 5
| Band | Observable qualities | What to listen / look for |
|---|---|---|
| 1 | Frequent sound / stress errors | About a third may be hard to identify |
| 3 | Most sounds correct | A few words may remain unclear |
| 5 | Readily understood sounds | Appropriate word / sentence stress |
Plain-language summary · Pearson PTE Core Score Guide, p. 39
Explanation
Fluency is about the flow. Pronunciation addresses the words within that flow. At pronunciation band one, frequent sound and stress errors may make about a third of the words difficult to understand. At band three, most sounds are correct, although recurring errors can leave a few words unclear. At band five, sounds are readily understood and word and sentence stress are appropriate.
That gives us a concrete practice target: make the important information recoverable on the first listen. For this chart, that means clear transport labels and clear numbers. We cannot judge pronunciation from the written answer. We need a recording and a listener who can identify what was actually said.
Technique 5: make the key words recoverable
Cars account for forty percent.
| Phrase | Role |
|---|---|
| fourTEEN | Different number |
| FORty | Target number |
| forty percent | Restore the unit |
Typical stress in isolated words; keep the number and unit recognisable in speech.
Explanation
Technique five is to rehearse the information-bearing word, then restore its phrase. Here the intended number is forty. Compare it with fourteen.
Response model
fourteen
Response model
forty
Response model
Cars account for forty percent.
The capitals show the usual stress when these words are said on their own. In a sentence, emphasis also depends on the message. If a listener hears fourteen, inspect the sound and stress contrast. If the speaker clearly chose the wrong number after misreading the chart, inspect Content. A pronunciation difficulty can also make a figure inaccurate in the message, but no fixed double deduction is published. Judge the actual recording, not its transcript.
The unit must remain audible too
Library graph: 2024 value = 50; unit = thousands of visits.
Wrong quantity
Visits reached fifty.
Number + unit
Visits reached fifty thousand.
Key point: Read the unit, then make it audible in the phrase.
Scoring connection: Content and Pronunciation: diagnose the actual cause
Explanation
The unit also carries information. Our library graph labels visits in thousands. At the 2024 point, the displayed value is fifty. Hear the difference.
Response model
Visits reached fifty.
Response model
Visits reached fifty thousand.
The first describes a different quantity. Rehearse the complete group.
Response model
fifty thousand visits
If thousand was intended but unclear, practise the phrase. If the unit label was misread, correct the interpretation first. Sound practice alone cannot fix the wrong quantity.
Preparation and transfer
A preparation routine that serves all three traits
- Identify the measure
- Select and connect the information
- Settle the opening and key words
Explanation
Here is how I use the preparation time. First I identify the topic and unit, so I know what the figures refer to. Then I mark the main feature, related details and a supported conclusion. Finally, I settle the opening phrase and any number I might stumble over. These are preparation decisions, not sentences I must write out. On this chart my notes are: commuting shares, car forty, bus thirty plus train twenty, bicycle ten.
With that route settled, I can devote more attention to speaking in complete phrases when the microphone opens.
The same method adapts to a different image
- Collection
- Sorting
- Delivery
This diagram has no branch; do not invent an address problem.
Explanation
We have seen why the relationship must follow the image. On this simpler parcel process, collection leads to sorting and then delivery. Unlike the earlier branched process, no alternative route is shown. We can describe the sequence and finish with delivery as the outcome. Adding an address problem here would invent a stage. The same content principle governs both diagrams: cover the visible stages and explain their connections.
Now we will hear three complete responses, beginning with the commuting chart, then a trend graph, and finally the branched process.
Complete demonstrations
A complete response using the method
Main commuting mode
| Category | Value |
|---|---|
| Cars | 40 |
| Buses | 30 |
| Trains | 20 |
| Bicycles | 10 |

Respondents (%) · One main mode per person · Complete model
Explanation
Response model
The chart shows the percentage of respondents using each main commuting mode. Cars are the largest individual category at forty percent. Buses account for thirty percent and trains for twenty percent, giving public transport a combined share of fifty percent. Bicycles are the smallest category at ten percent. Overall, public transport exceeds cars by ten percentage points, even though cars lead as a single category.
A complete response to the trend graph
Annual library visits
| Category | Value |
|---|---|
| 2021 | 20 |
| 2022 | 40 |
| 2023 | 30 |
| 2024 | 50 |

Visits (thousands) · Overall increase; a decline in 2023
Explanation
Response model
The line graph shows annual library visits in thousands from twenty twenty-one to twenty twenty-four. Visits rose from twenty thousand to forty thousand in the first year, then fell to thirty thousand. They recovered to fifty thousand in the final year, the highest figure shown. Overall, visits increased across the period, although growth was interrupted by a decline in twenty twenty-three.
A complete response to the branched process
Starting sequence: Collection<br>then sorting
Decision: Address valid?
| Condition | Next step |
|---|---|
| Yes | Delivery |
| No | Return to sender |

The address check determines two possible outcomes.
Explanation
Response model
The diagram shows how parcels move through collection, sorting and an address check. After collection, parcels enter sorting, where the address determines their route. Those with a valid address proceed to delivery. Those with an invalid address return to the sender. The process therefore has two possible outcomes, rather than successful delivery for every parcel.
What changed across the three complete answers
| Image | Relationship to develop | Supported ending |
|---|---|---|
| Commuting shares | Category comparison + grouping | 50% versus 40% |
| Annual visits | Direction + turning point | Growth despite a decline |
| Parcel process | Stages + condition + routes | Two possible outcomes |
Reuse the reasoning; adapt the description to the actual image.
Explanation
Notice what changed across those three answers. The commuting chart needed category comparisons and a compatible grouping. The line graph needed direction and a turning point. The process needed stages, a condition and two outcomes. Each answer developed relationships, but the relationship was different. That is how to reuse a technique without forcing a memorised sentence onto a new image. Before speaking, identify the kind of connection the image actually supports.
Then choose wording you can deliver smoothly and accurately. The scoring purpose stays consistent while the description changes.
Review
Review the performance by trait
| Criterion / example | Meaning / explanation |
|---|---|
| Content | Coverage + relationships + supported conclusion |
| Oral Fluency | Continuity + natural phrasing |
| Pronunciation | Recognisable words + appropriate stress |
Diagnose an audible or visible weakness, revise it, then compare the recording.
Explanation
The commuting answer covers all four categories and develops the fifty-versus-forty comparison. That addresses the content weakness in the original list. Across all three models, the sentences follow a planned route, which helps keep the delivery organised. To assess your own attempt, listen once for the information and once for the speech. Check for missing features or unsupported claims, then for restarts and unclear words.
Choose the next revision from the weakness you actually hear. These techniques help you build the qualities described by the rubric; no particular phrase can guarantee a score. You now have a method for improving the answer and checking whether the improvement worked.