The interquartile range (IQR) tells you how spread out the middle 50% of your data is. A small IQR means the central values are tightly clustered; a large IQR means they’re widely scattered. Once you understand what that number represents and how to use it, you can quickly judge variability, spot outliers, and compare datasets at a glance.
What the IQR Actually Measures
To calculate the IQR, you first sort your data from lowest to highest. Then you find two boundary points: the first quartile (Q1), the value below which 25% of your data falls, and the third quartile (Q3), the value below which 75% of your data falls. The IQR is simply Q3 minus Q1.
That single number captures the range occupied by the middle half of your dataset, ignoring the lowest 25% and the highest 25%. This makes it naturally resistant to extreme values. One wildly high or low measurement won’t drag the IQR the way it would drag a standard deviation, which is why it’s the go-to measure of spread whenever data is lopsided or contains unusual values.
What a Large or Small IQR Tells You
The size of the IQR only means something relative to the scale you’re measuring. On a 100-point math exam, an IQR of 2 means the middle half of students scored within just 2 points of each other. That’s remarkably consistent performance. An IQR of 18 on the same exam tells you students in the middle group varied by nearly a fifth of the total scale, a much more mixed picture.
So when you see an IQR, ask two questions: How big is it compared to the full range of possible values? And how big is it compared to the IQR of another group I’m comparing against? A narrow IQR signals agreement or consistency. A wide one signals diversity or inconsistency. Neither is inherently good or bad; it depends on context. Tight clustering in manufacturing quality data is usually desirable. Tight clustering in creative test scores might suggest the test isn’t distinguishing between students well enough.
How to Calculate It Step by Step
Start by ordering your data from smallest to largest. Find the median, which splits the dataset into a lower half and an upper half. Q1 is the median of the lower half, and Q3 is the median of the upper half. Subtract Q1 from Q3, and you have the IQR.
A quick example: suppose you have the values 3, 5, 7, 8, 12, 14, 16. The median is 8 (the middle value). The lower half is 3, 5, 7, so Q1 is 5. The upper half is 12, 14, 16, so Q3 is 14. The IQR is 14 minus 5, which equals 9. That means the middle 50% of your data spans 9 units.
When your dataset has an even number of values, the median falls between two middle numbers (their average), and you split the data evenly into lower and upper halves before finding Q1 and Q3 the same way.
Using the IQR to Spot Outliers
One of the most practical uses of the IQR is flagging data points that fall unusually far from the pack. The standard rule, introduced by the statistician John Tukey in 1977, works like this: multiply the IQR by 1.5. Any value more than 1.5 × IQR below Q1 or above Q3 is considered an outlier.
These boundaries are called “inner fences.” To find them:
- Lower fence: Q1 minus 1.5 × IQR
- Upper fence: Q3 plus 1.5 × IQR
Anything outside those fences is a mild outlier. Tukey also defined “outer fences” at 3 × IQR beyond the quartiles. Data points past the outer fences are considered extreme outliers, values so far from the center that they almost certainly reflect errors, unusual circumstances, or a fundamentally different process.
For the example above (Q1 = 5, Q3 = 14, IQR = 9), the lower fence would be 5 minus 13.5, which is −8.5, and the upper fence would be 14 plus 13.5, which is 27.5. Any value below −8.5 or above 27.5 would be flagged. In this small dataset nothing qualifies, but in larger, messier datasets this rule quickly isolates the values worth investigating.
Reading the IQR on a Box Plot
If you’ve seen a box-and-whisker plot, you’ve already seen the IQR. The box itself represents the middle 50% of the data. Its left (or bottom) edge sits at Q1, its right (or top) edge sits at Q3, and the line inside the box marks the median. The length of the box is the IQR.
The “whiskers” extend from the box to the smallest and largest values that still fall within the inner fences. Any points plotted individually beyond the whiskers are outliers. When you’re comparing two box plots side by side, the relative lengths of the boxes immediately tell you which group has more variability in its core data. A short box means consensus; a long box means spread.
The position of the median line inside the box also reveals skew. If the median sits closer to Q1 with more space stretching toward Q3, the data skews toward higher values. If it sits closer to Q3, the data skews lower.
Why You’ll See IQR Instead of Standard Deviation
Standard deviation is the most common measure of spread, but it assumes your data is roughly symmetrical. When data is skewed, meaning it bunches up on one side with a long tail stretching the other way, the mean gets pulled toward the tail and the standard deviation inflates. Neither number faithfully represents where most of the data actually sits.
In those situations, the median paired with the IQR gives a more honest picture. In a study of hemoglobin levels in 48 intensive care patients, for instance, the full range spanned from quite low to quite high values, but the IQR showed that the central 50% of patients fell within a much narrower band of 8.7 to 10.8 g/dl. That tighter window was far more useful for understanding typical patient values than the overall range, which was stretched by a few extreme readings.
You’ll commonly encounter IQR reporting in health research, salary data, real estate prices, hospital length-of-stay figures, and any other domain where a handful of extreme cases can distort averages. When a study reports something like “median age 54 years (IQR 47 to 63),” it’s telling you the middle value and the range that captured the central half of participants, giving you both a best guess and a quick sense of variability in two numbers.
Quick Interpretation Checklist
When you encounter an IQR in a report, article, or your own analysis, run through these points:
- Scale matters. Compare the IQR to the full range of the measurement. An IQR of 10 means something very different on a 20-point scale versus a 1,000-point scale.
- Compare groups. The IQR is most informative when you’re comparing two or more datasets. A wider IQR in one group signals more variability in that group’s middle 50%.
- Check for outliers. Apply the 1.5 × IQR rule to identify mild outliers and the 3 × IQR rule for extreme ones.
- Pair it with the median. The IQR describes spread; the median describes the center. Together they give you a resistant summary of your data that won’t be thrown off by a few unusual values.
- Consider the shape. If Q3 is much farther from the median than Q1, the upper half of your data is more spread out, suggesting a right-skewed distribution.