A Cleveland dot plot displays one dot per category along a horizontal axis, replacing bar chart bars with dots. A dumbbell chart (or connected dot plot) adds a second dot per category and connects the two with a line — showing how a value changed between two points in time or between two groups.
You can paste two values per category into the free Cleveland dot plot maker to create either a single-dot comparison or a connected dumbbell chart.
The core idea
William S. Cleveland introduced the dot plot in his 1984 research as a more perceptually accurate alternative to the bar chart. The insight: humans judge positions on a common scale more accurately than they judge the height of bars rooted at a shared baseline. Removing the bar and showing just the dot reduces visual clutter and shifts the reader's eye to what matters — where the dot lands.
The dumbbell chart extends this to two series per category. Each category gets a dot for Group A and a dot for Group B, connected by a line segment. The line makes the gap between the two values immediately visible — and the direction of the gap (left vs right) shows which group is higher.
Single dot plot vs dumbbell chart
- Single Cleveland dot plot — one dot per category. Functions like a horizontal bar chart, often used to rank categories. Cleaner than bars when many categories are listed vertically.
- Dumbbell chart (connected dot plot) — two dots per category, connected by a line. Used to compare two groups or two time points (Before vs After, 2020 vs 2024, Men vs Women).
What dumbbell charts reveal
- Line length → the magnitude of change or difference between two groups
- Line direction → which group is higher (dot on the right = higher value)
- Sorted order → sorting categories by line length ranks them from biggest to smallest gap
- Color coding → each dot color identifies which group it belongs to
A dumbbell chart is the fastest way to answer "Which categories changed the most?" — something bar charts require side-by-side scanning to answer.
Cleveland dot plot vs bar chart
| Cleveland dot plot | Bar chart | |
|---|---|---|
| Perceptual accuracy | Higher (position judgment) | Slightly lower (length judgment) |
| Visual clutter | Low (no filled bars) | Moderate |
| Two groups per category | Natural (dumbbell) | Grouped bars (crowded) |
| Baseline at zero required | No | Yes (misleading if not) |
| Familiar to general audiences | Less familiar | Very familiar |
The main trade-off: bar charts are universally recognized; dot plots are less common and may need a brief explanation for non-technical audiences.
When to use a Cleveland dot plot
- Ranking many categories (15–30+ items) where bars would be too wide or too narrow
- Before/after comparisons: policy changes, A/B tests, year-over-year shifts
- Comparing two groups across multiple categories (e.g., male vs female income by city)
- Any situation where the gap between two values is the story
FAQ
Is a dumbbell chart the same as a connected dot plot?
Yes — "dumbbell chart," "connected dot plot," and "gap chart" all refer to the same thing: two dots per row connected by a horizontal line. The name "dumbbell" comes from the visual resemblance to a dumbbell weight when the two dots are large.
Why sort by difference in a Cleveland dot plot?
Sorting categories by the gap between the two dots (Value2 − Value1) turns the chart into a ranking of change magnitude. The categories with the biggest shifts appear at the top (or bottom), making the most important patterns immediately visible without scanning the whole chart.
How is a Cleveland dot plot different from a scatter plot?
In a scatter plot, each dot represents a single observation with two measured variables (X and Y). In a Cleveland dot plot, each row is a named category and the horizontal axis shows one (or two) values for that category. The axes are different: scatter plots have two numeric axes; dot plots have one numeric axis and one categorical axis.
How this guide is reviewed
This guide is maintained by the ChartsMakers Editorial Team and reviewed for statistical accuracy, product behavior, and clarity. When a guide relies on software behavior or statistical definitions, we check it against official documentation or established references. See our editorial policy.