Bar Graph in IELTS Task 1: The Three Shapes, and Where the Marks Actually Go

A bar graph is the most common question type in IELTS Academic Writing Task 1, and almost every weak answer fails in the same place. Not in the numbers. In the decision about which numbers to put next to each other.

This post does three things. It separates the three shapes a bar graph actually arrives in, because each one asks for a different grouping. It walks through the four paragraphs. Then it takes an answer published in the Bandly sample library and shows where its marks went, criterion by criterion.

What Task 1 is asking for

In Academic Writing Task 1 you get a visual and 20 minutes. You must write at least 150 words. The instruction is always close to this wording:

Summarise the information by selecting and reporting the main features, and make comparisons where relevant.

Three words in that sentence decide your mark. Summarise means do not write everything. Selecting means you choose, and the choice is marked. Comparisons means the examiner wants figures held against each other, not a list.

You are not asked for an opinion, a reason, or a conclusion. A bar graph showing that car use is high in Houston does not want your view on cars. If you explain why the data looks like this, you are writing content the task did not ask for, and that costs you under the first criterion.

Your answer is marked on four criteria. In Task 1 they are Task Achievement (did you cover and summarise the right things), Coherence and Cohesion (is it organised and linked), Lexical Resource (vocabulary), and Grammatical Range and Accuracy (sentence structure and errors). Each one is marked separately, and your Task 1 answer counts for about one third of your writing band. The other two thirds come from Task 2.

The three shapes a bar graph arrives in

Most guides treat "bar chart" as one question type. It is not. The horizontal axis tells you which of three jobs you have, and the job changes what you group.

Categories only. There is no time, so there is no trend. Words like "increased" and "fell" do not belong here. You are ranking and contrasting. This is also the shape where a static chart with no time axis catches candidates who have memorised trend language and nothing else.

Time series bars. The bars behave exactly like a line. You describe movement, so you need rises, falls, peaks and the steepest change. Our guide to describing trends in Task 1 covers the grammar that goes with those words, which is where most of the lost marks sit.

Two charts together. A bar chart beside a table or a pie chart is one question, not two. You still write one overview covering both. The way to handle a Task 1 question with two charts is to give each visual its own body paragraph and to point at the relationship once in the overview.

Read the axis labels and the key before you write a single word. Thirty seconds spent on the unit of measurement, whether it is a percentage, thousands of people or millions of dollars, saves you from an answer that is well written and reports the wrong thing.

The four paragraphs

Every bar graph answer, in any of the three shapes, fits this shape.

1. Introduction, one sentence. Say what the chart shows, in your own words. Change the wording of the question. Do not copy it, because copied wording is not counted in your word total.
2. Overview, one or two sentences. The biggest pattern on the page, with no individual figures. This paragraph does more for your band than any other.
3. Body paragraph one. One group of data, with figures.
4. Body paragraph two. The other group, with figures, compared to the first.

The overview is the paragraph candidates skip when they run short of time, and it is the most expensive thing to skip. The public band descriptors for Task 1 expect a clear overview of the main trends or differences, and an answer with no overview is capped low on Task Achievement no matter how good the rest of the English is. If you only have two minutes left, write the overview and leave a body paragraph short. We go through that single sentence in detail in the Task 1 overview guide.

Grouping is the whole skill

Here is the difference between a band 6 answer and a band 7 or 8 answer on the same chart. It is not vocabulary. It is grouping.

A weak answer walks the chart from left to right and reports every bar in the order it appears. Tokyo train, Tokyo bus, Tokyo car, Tokyo bicycle, then Amsterdam, then Houston. Twelve figures, twelve clauses, no comparison. This reads as a list, and the examiner cannot see any selection happening.

A stronger answer puts figures together for a reason. Three groupings usually work:

• By extreme. The highest and the lowest, in one sentence, with the gap between them.
• By similarity. Two cities or two products that behave the same way, in one clause.
• By exception. The group that breaks the pattern the overview just described.

The test is simple. Read your body paragraphs and count the sentences that contain two figures. If every sentence has exactly one figure, you have listed rather than compared, and no amount of better vocabulary will fix it. Choosing which features to report is the part of this task that is actually being marked.

A worked answer, and where its marks went

The sample library holds a bar chart answer to this question: the percentage of commuters using four types of transport in three cities in 2023. You can read the whole thing, with every scored sentence highlighted, on the bar chart samples page. The band estimates there come from Bandly's own engine, not from a qualified examiner, so treat them as estimates and read the reasoning rather than the number.

Its overview is this:

Overall, there are stark contrasts in commuting preferences among the three urban areas. Tokyo is heavily dominated by train travel, Amsterdam shows a strong preference for cycling, and Houston is overwhelmingly reliant on cars, while bus travel remains relatively unpopular in all three locations.

Two sentences, no figures, and three groupings plus one thing they share. That is the pattern to copy.

The answer was estimated at 7.5 overall, and the four criteria did not agree with each other:

Read the right hand column again. Nothing there is about the chart. The data was handled well. What held the answer at 7.5 was the language wrapped around the numbers.

What the published answers lost their marks on

I counted this across the whole Task 1 Academic set in the library, which is four published answers at the time of writing: one bar chart, one line graph, one table and one process diagram. Each published answer carries a band estimate per criterion. Here is every one of them.

Task Achievement came back 8 or 9 on all four. Coherence and Cohesion and Lexical Resource came back 6 or 7 on all four. Not one of these answers was held back by picking the wrong data.

Two honest warnings about that count. Four answers is a small number, and these are machine estimates that no qualified examiner has checked, so the right reading is a direction, not a measurement. But the direction matches what the feedback text actually says. The repeated complaints are a thin set of linking words, a repeated phrase, a wrong preposition, a singular where a plural belongs.

The practical consequence is this. If you can already find the main pattern in a chart, more chart practice will not move your band. Fixing the phrase you repeat four times will.

The language that does the comparing

Two patterns carry most of a bar graph answer.

Comparing two figures in one clause.

• Train use in Tokyo, at 62%, was five times the figure for bicycles.
• Bus travel was the least popular option in all three cities, at between 5% and 14%.
• Car use in Houston reached 88%, compared with 25% in Amsterdam.

Naming an extreme.

• The most common mode in Amsterdam was cycling, chosen by 48% of commuters.
• Trains accounted for the smallest share in Houston, at just 3%.

Keep the numbers themselves plain. Percentages stay as figures with the symbol. Whether you write a figure or spell a number out is covered in how to write numbers in Task 1, and getting it wrong is a small but repeated Lexical Resource error in the samples above.

Avoid three habits that show up constantly in weak answers. Do not write "respectively" with a list longer than three items, because the reader loses the mapping. Do not start three sentences in a row with the same city name. Do not use "dramatically" for a difference between categories, since nothing is moving.

What to do on your next bar graph

1. Name the shape first. Categories, time, or two charts.
2. Write the overview before the body paragraphs, not after.
3. Group your figures, then count the sentences with two numbers in them. Fewer than two per paragraph means you listed.
4. Pick the phrase you repeated most and replace it once.

Do this on one chart rather than five. When you have a finished answer, read the scored Task 1 Academic samples beside it and compare the sentences, not the bands. If you want a second read on your own answer, you can paste it into Bandly and get a band estimate for each of the four criteria with the sentences that drove it marked.