Dot Plot Explained: Meaning, Interpretation and How to Make One in SPSS, R and Python
A dot plot shows every single observation as a dot, stacked where values are close together. This guide explains what a dot plot is, when a dot plot is better than a bar chart or histogram, how to read clusters, gaps and overlap, and how to create a dot plot in SPSS, R, Python or free online.
Paste grouped data and get a dot plot with mean and 95% CI, rank-based tests and APA 7 text.
Open the free dot plot calculator
What is a dot plot?
A dot plot (here the stacked dot plot of Wilkinson) draws one dot per observation along a numeric axis. Values that are equal or fall in the same narrow bin stack on top of each other, so the height of a stack shows frequency, much like a histogram, while every individual value stays visible. Not to be confused with the Cleveland dot plot, which displays one summary value per category. A grouped dot plot gives each group a separate row on a shared axis for direct comparison.
When should you use a dot plot?
- Small and moderate samples (roughly 5 to 100 per group), where a histogram would show only a few lumpy bars.
- Discrete or rounded data such as counts, Likert scores, ages in years or scores out of 10.
- Pilot studies and teaching, where the reader should see every data point.
- Replacing bar charts of means. A dot plot reveals spread, overlap and outliers that a bar with an error bar hides.
- Before choosing a test. Skew, gaps and outliers in the dot plot argue for rank-based methods.
Why choose a dot plot over a bar chart or histogram?
A bar chart of means reduces each group to two numbers and can make very different distributions look identical. A histogram needs enough data to fill the bins. A dot plot is transparent: the reader sees the sample size, the spread, clusters, gaps and outliers immediately, and no information is lost. Adding a mean with a 95% confidence interval on the same dot plot gives both the raw data and the summary.
How to interpret a dot plot
| Feature of the dot plot | Interpretation |
|---|---|
| Tallest stack | Most frequent value or bin (the mode) |
| Horizontal extent of the dots | Range and spread of the group |
| Dots piled toward one end with a thin tail | Skewed distribution; the median describes the centre better than the mean |
| Isolated dot far from the rest | Possible outlier; check data entry before removing |
| Gap between clusters | Possible subgroups or measurement artefacts |
| Heavy overlap of groups | Weak separation, even when a test is significant in a large sample |
| Diamond with bar | Group mean with the 95% confidence interval |
The bin width controls how dots group: very narrow bins spread dots out until stacks disappear, very wide bins merge distinct values. For small samples, prefer rank-based tests (Mann–Whitney U or Kruskal–Wallis with Dunn post hoc tests) and report medians with the interquartile range.
Common mistakes with a dot plot
- Using a dot plot with hundreds of observations per group, which overloads the figure; use a density plot instead.
- Choosing a bin width that is much smaller or larger than the measurement resolution.
- Confusing the stacked dot plot with the Cleveland dot plot of category summaries.
- Reading the vertical stack as a measured variable; the stack only shows frequency.
- Showing only the mean and ignoring the spread of the dots.
How to report a dot plot in APA 7
Example: Individual observations were displayed in a dot plot (Figure 1). A Kruskal–Wallis test indicated no significant difference among groups, H(2) = 2.45, p = .293, ε² = .12. Put medians and interquartile ranges in a table and describe the dot plot in the figure note.
How to make a dot plot in SPSS, R and Python
SPSS
- Choose Graphs → Chart Builder.
- In the Gallery pick Scatter/Dot and drag Simple Dot Plot to the canvas.
- Drag the numeric variable to the x-axis and the grouping variable to Panel by (rows) so each group gets a separate stacked row.
- Click OK, then double-click the chart to adjust the bin width and dot size.
For group tests choose Analyze → Nonparametric Tests → Independent Samples; SPSS runs Mann–Whitney U for two groups or Kruskal–Wallis with Dunn–Bonferroni pairwise comparisons for three or more.
R
library(ggplot2)
ggplot(df, aes(x = Value)) +
geom_dotplot(binwidth = 0.5) +
facet_grid(Group ~ .)Python
import seaborn as sns
sns.stripplot(data=df, x="Value", y="Group", size=8)
# or sns.swarmplot for non-overlapping dotsFull scripts with tests are in the dot plot tutorial.
The easy way: dot plot on StatFAS.com
The free dot plot calculator draws a stacked dot plot per group with an optional mean and 95% CI marker, reports descriptives and Shapiro–Wilk, runs Mann–Whitney or Kruskal–Wallis with Dunn–Holm post hoc tests, and writes the APA 7 results. The number of bins, colours, fonts and labels are editable, and signed-in visitors can download a combined Word report.
Paste grouped data and get a dot plot with mean and 95% CI, rank-based tests and APA 7 text.
Open the free dot plot calculator
Frequently asked questions about the dot plot
What does a dot plot show?
A dot plot shows each observation as a dot on a numeric axis, stacking dots with similar values so that frequency, spread, clusters and outliers are visible.
What is the difference between a dot plot and a histogram?
A histogram shows counts in bars and hides individual values, while a dot plot keeps every observation visible, which suits small samples.
When is a dot plot better than a bar chart?
A dot plot is better when the reader needs to see the sample size, spread and outliers, because a bar chart of means hides all three.
Which test goes with a dot plot of three groups?
A Kruskal–Wallis test followed by Dunn post hoc tests with Holm correction is a standard rank-based choice for small, skewed samples.
How many observations can a dot plot handle?
A dot plot stays readable up to roughly 100 observations per group; beyond that a density plot or box plot is clearer.