Worked examples
- Box plot example: plant growth
- Box plot example: laboratory measurements
- Box plot example: ties and outliers
How to make a box plot
- Paste or open a raw-observation table with one group per column and headers in the first row.
- Select group columns, inspect sample counts and excluded cells, then confirm the input format.
- Choose orientation and All observations or Outliers only; the calculations never change.
- Inspect the summary table, set journal sizing, save a project and export.
Data format and interpretation
Each numeric cell is one observation; group sizes may differ. Blank cells and ordinary missing markers such as NA are excluded with counts. Invalid text and Excel errors block export. Do not paste means, frequencies or five-number summaries as observations. The maximum is 20 groups and 3,500 accepted observations; no sampling occurs.
| Control | Treatment |
|---|---|
| 1 | 2 |
| 2 | 3 |
| 3 | 4 |
| 4 | 5 |
| 10 | NA |
The box uses type-7 quartiles. Whiskers end at actual observations inside the fences, following the documented boxplot convention. Outliers always remain visible and in the numeric range. A zero-IQR box is a line, not evidence of data removal.
Small samples provide limited information about a distribution. Deterministic point offsets separate observations along the group direction only; the measured value never changes. Dense groups may still overlap, so inspect n and the numeric tables.
Frequently asked questions
What do the box and whiskers show?
The box spans type-7 Q1 to Q3, with a median line. Whiskers reach the most extreme observed values within inclusive fences Q1−1.5×IQR and Q3+1.5×IQR. They are not confidence intervals.
Why do quartiles differ from another program?
Quartile algorithms differ, especially for small samples. This tool uses R quantile type 7, shared with Scientific Plot Maker. Default R boxplot hinges are not identical for every sample size.
Does an outlier mean a data error?
No. Here it means a value outside the 1.5-IQR fences. Outliers remain observations in every calculation and axis range. The tool never removes them or declares them mistakes.
What happens to small or empty groups?
Empty groups retain a No values slot. A singleton shows its point without a box. With two observations, quartiles use interpolation. Zero-IQR boxes collapse to a line; unequal observations can still be outliers.
Can I paste a five-number summary?
No. Enter individual numeric observations, one group per column. The editor cannot reliably identify numeric summary rows, so you must confirm the selected columns contain raw observations.
Are group differences statistically significant?
This tool does not test significance or calculate confidence intervals. Box overlap and outlier status are descriptive, not inferential conclusions.
Inspect one distribution with the Histogram Maker, compare supplied category values with the Bar and Column Chart Maker, or explore other descriptive summaries in Scientific Plot Maker.