Double bar graphs — comparing two sets
A double bar graph is like an ordinary bar chart, but for each group on the x-axis there are two bars side by side, one for each series. That makes it perfect for comparing — same group, two different things.
Some real examples:
- Number of boys vs girls in each class.
- Goals scored vs goals conceded by each football team.
- Sunny days vs rainy days in each month.
- This year's sales vs last year's sales for each shop.
How to read one
- Look at the legend (key). It tells you which colour is which series.
- For one bar: find the top of the bar and read across to the y-axis.
- For one group across two series: read both bars in that group. Compare or subtract.
- For trends across groups: scan the same colour across all groups — that's how the first series changes from group to group.
Worked example
The graph above shows boys and girls in four Year 5 classes.
Q1: How many girls are in class 5B?
Find class 5B. The orange bar (girls) reaches up to 20. 20 girls.
Q2: How many more girls than boys are there in class 5B?
Boys in 5B: 16. Girls in 5B: 20. 20 − 16 = 4 more girls.
Q3: Which class has the largest difference between boys and girls?
Scan each pair:
- 5A: 24 boys, 18 girls → diff 6 (boys more).
- 5B: 16 boys, 20 girls → diff 4 (girls more).
- 5C: 22 boys, 12 girls → diff 10 (boys more).
- 5D: 14 boys, 18 girls → diff 4 (girls more).
Largest difference is 5C at 10.
Tips
- Don't compare different-colour bars across groups. "Boys in 5A vs girls in 5C" doesn't usually make sense as a comparison — you're mixing two unrelated things.
- Same-colour bars across groups are comparable. "Boys in 5A vs boys in 5B" is fine.
- Same group, two colours is the main use. That's what double bar graphs are designed for.
- Watch the scale. As with line graphs, a squashed scale can make a small difference look big.
When to choose a double bar graph
- You have two related sets to compare across the same groups. ✅
- You have three or more sets: use a triple bar graph or a grouped bar chart — same idea, more bars per group.
- You have time-series data: use a line graph, not bars.
- You have categorical data with no natural pairing: a single bar chart is enough.
A puzzle
A shop tracks ice creams and hot drinks sold each season. Ice creams: spring 30, summer 80, autumn 25, winter 5. Hot drinks: spring 40, summer 10, autumn 60, winter 90.
- In which season is the largest difference between the two products?
Differences: spring 10, summer 70 (ice creams win), autumn 35, winter 85 (hot drinks win).
Largest is winter at 85. (Hot drinks crush in winter, as expected.)