Reading Charts, Tables and Figures
How to read a graph in English, describe what it shows, and spot the presentation tricks that make data say more than it should.
Charts appear in exams, reports and news. Reading them needs two things: the language of change, and a suspicion of the axes.
Read the furniture first
Before the shape of the line, read the labels around it — the four things that tell you what you are looking at:
- The title — what is being measured
- The axes — what each direction represents, and in what units
- The legend — which line or colour is which
- The footnote — the sample, the source, and any exclusions
The footnote is where the qualifications live: provisional, excluding Scotland, self-reported, n=112. It is small on purpose and it is often the most important line on the chart.
Which chart for what
| Type | Shows |
|---|---|
| line graph | change over time |
| bar chart | comparison between categories |
| pie chart | parts of one whole (should total 100%) |
| scatter plot | relationship between two variables |
| table | exact values |
A pie chart whose segments don't add to 100%, or a line graph of unrelated categories, is a sign of carelessness at best.
The language of change
Verbs up: rise · increase · grow · climb · surge · peak Verbs down: fall · drop · decline · decrease · plummet · bottom out No change: remain stable · level off · plateau · hold steady Movement: fluctuate (up and down) · recover · double · halve
The grammar trap: rise and fall take no object; raise and lower need one.
✅ Sales rose. · ✅ They raised prices.
❌ Sales raised.
Describing the size of a change
Adverbs: sharply · steadily · gradually · slightly · dramatically · marginally Adjectives: a sharp rise · a steady decline · a slight increase
Revenue rose sharply in Q1, then levelled off.
There was a gradual decline over five years.
Note both patterns: verb + adverb and adjective + noun. Exams reward using both.
Comparing
- twice as many as · three times higher than · half as much as
- compared with / in contrast to · by comparison
- the gap widened / narrowed
- X outperformed Y · X lagged behind Y
- per head / per 1,000 people — a rate, not a total
That last one matters: a country with more cases may have fewer cases per head. Totals and rates routinely tell opposite stories, and switching between them is a standard way to mislead honestly.
Precision language
- approximately / roughly 40%
- just over / just under half
- around a third
- the vast majority (well over half)
- a significant minority
What to check before believing a chart
1. Does the y-axis start at zero? If not, small differences look enormous. This is the most common way an honest number is made to look dramatic.
2. What is the time range? A five-year window can show the opposite trend to a fifty-year one.
3. Are the intervals even? Uneven spacing on the x-axis distorts the shape of the line.
4. Percentage or absolute? "Crime doubled" from 2 cases to 4 is true and almost meaningless.
5. What is the sample? A survey of readers of one newspaper is not a survey of the public.
6. Correlation or causation? Two lines moving together prove nothing about one causing the other.
7. Percentage of what? A rise "from 3% to 6%" can be reported as up 3 percentage points or a 100% increase. Both are true; they feel completely different.
Describing without over-claiming
✅ The data show a correlation between X and Y.
⚠️ The data prove X causes Y.
Examiners and careful readers both notice this distinction.
A worked description
The chart shows monthly sales from January to December. Sales rose steadily in the first quarter, peaking in April at just over £40,000. There was then a sharp decline to a low of £18,000 in July, before figures recovered gradually over the autumn. Overall, sales ended the year slightly higher than they began.
Notice the shape: what it shows → the movements in order → the extremes with figures → an overall summary. That structure is what exam mark schemes reward, and it works for any chart.
The chart is an argument. Someone chose the time range, the axis and the comparison, and each choice makes some conclusions visible and others invisible. Asking what would this look like drawn differently? is not cynicism — it is the same question the person who made it had to answer.
Test yourself
The quiz below covers the verbs and the axis traps.