A box plot packs five statistics and an outlier rule into one small shape. This guide walks through reading one from the inside out — the same order experienced analysts use.
Step 1: Find the Median Line
The line inside the box is the median — the 50th percentile. It is the single best one-number summary here because, unlike the mean, it is not dragged around by extreme values. When comparing several boxes, start by comparing median heights.
Step 2: Judge Spread From the Box
The box runs from Q1 to Q3, so its length is the IQR — the range covered by the middle half of the data. A short box means consistent, tightly clustered values; a long box means high variability. Two groups can share a median while one is twice as spread out, so never compare medians alone.
Step 3: Read Skewness From Asymmetry
- Median centered, whiskers equal — roughly symmetric data.
- Median near Q1, long upper whisker — right skew (income, reaction times).
- Median near Q3, long lower whisker — left skew (easy-exam scores).
Step 4: Check the Outlier Dots
Dots beyond the whiskers are values outside the Tukey fences (Q1 − 1.5×IQR, Q3 + 1.5×IQR). They are not automatically errors — they are candidates for investigation. A cluster of outlier dots on one side usually means the whole distribution is skewed, not that many points are broken.
Step 5: Compare Groups Correctly
With multiple boxes, look for three things in order: do the medians differ? Do the boxes (IQRs) differ in length? Do the boxes overlap? Non-overlapping notches (available in this site's box plot tool) give a quick visual significance check around the medians.
What a Box Plot Cannot Tell You
A box plot hides modality and detail inside the quartiles. Two datasets with identical five-number summaries can have completely different shapes — one a clean bell curve, the other two distinct clusters. If shape matters, verify with a violin plot or histogram of the same data.