Yes, the interquartile range (IQR) is resistant to outliers. Because it measures only the spread of the middle 50% of your data, extreme values at either end have no effect on it. This makes the IQR one of the most commonly recommended measures of spread when your dataset contains unusual or extreme observations.
Why the IQR Ignores Extreme Values
The IQR is calculated by subtracting the 25th percentile (Q1) from the 75th percentile (Q3). That means it only cares about where the middle half of your data falls. If you have a dataset of test scores where most students scored between 80 and 90, the IQR is 10, regardless of whether the lowest score was 74 or 12. The extreme value simply doesn’t enter the calculation.
Compare this to the range, which uses the single highest and single lowest values in the entire dataset. One wildly unusual observation can dramatically inflate the range while leaving the IQR completely unchanged. Khan Academy illustrates this with temperature data: a single low outlier in one location’s readings has a large impact on the range but zero impact on the IQR.
This property is called “resistance” or “robustness” in statistics. A resistant measure is one that doesn’t shift much when a few data points are replaced with extreme values. The IQR shares this property with the median, which is also resistant to outliers. The mean and standard deviation, by contrast, are not resistant because every data point pulls on the calculation.
How Much Data Can Be Extreme Before the IQR Breaks
A useful way to think about resistance is the concept of a breakdown point: the proportion of data that can be moved to infinity before a statistic becomes meaningless. The IQR has a breakdown point of 25%. You could replace up to a quarter of your data with absurdly extreme values and the IQR would still give a reasonable answer, because those values would shift only beyond Q1 or Q3 without changing the boundary between them.
The range, by contrast, has a breakdown point of 0%. A single extreme observation, no matter how large your dataset is, can distort the range. The standard deviation sits somewhere in between but is still highly sensitive to outliers because it squares each deviation from the mean, amplifying the influence of extreme values.
How the IQR Is Used to Detect Outliers
The IQR’s resistance to outliers is precisely what makes it useful for identifying them. The standard method, often called Tukey’s fences or the 1.5 IQR rule, works like this:
- Calculate the IQR: Q3 minus Q1.
- Build a lower fence: Q1 minus 1.5 times the IQR.
- Build an upper fence: Q3 plus 1.5 times the IQR.
- Flag outliers: Any data point below the lower fence or above the upper fence is considered an outlier.
For example, if Q1 is 80 and Q3 is 90, the IQR is 10. Multiply by 1.5 to get 15. The lower fence is 65 and the upper fence is 105. Any score below 65 or above 105 would be flagged. This approach works because the IQR itself isn’t distorted by the very outliers you’re trying to find. If you used a non-resistant measure to set your thresholds, extreme values would widen those thresholds and effectively hide themselves.
The CDC uses this same 1.5 IQR rule as its standard method for calculating outliers in data visualizations, though it notes the rule may not be appropriate for every purpose.
How Box Plots Make This Visible
Box-and-whisker plots are built directly around the IQR. The box spans from Q1 to Q3, giving you a visual snapshot of where the middle 50% of the data sits. The whiskers extend to the most extreme non-outlier values, and any points beyond the fences are plotted individually as dots or asterisks.
This design lets you instantly see two things: the typical spread of your data (the box) and how far any extreme points deviate from that spread (the isolated dots). Because the box is anchored to the resistant IQR, it stays stable even when outliers are present, giving you a reliable picture of the central tendency without being visually distorted by a few extreme observations.
When Resistance Matters Most
The IQR’s resistance is most valuable when you’re working with data that naturally produces extreme values: income distributions, home prices, medical lab results, or any dataset where a small number of observations can be orders of magnitude larger than the rest. In these situations, the standard deviation and range can give a misleading impression of how spread out the typical observations are. The IQR tells you how spread out the middle of the data is, which is often the more useful question.
If your data is clean and roughly symmetric with no outliers, the IQR and standard deviation will tell a similar story. The IQR’s advantage shows up specifically when the data is skewed or contains extreme values. In those cases, pairing the median with the IQR gives you a summary of center and spread that accurately represents what most of your data actually looks like.