How to Find the Range in Statistics and Identify Outliers

The range is one of the simplest ways to describe how spread out a data set is. It tells you the distance between the largest value and the smallest value in a list of numbers. The formula is short:

\( \)Range\( = \)Maximum value\( - \)Minimum value\( \)

Because it only uses two numbers, the range is quick to calculate, but it is also very sensitive to unusual values. That sensitivity is exactly why range and outliers are usually studied together.

To find the range, order the data from smallest to largest, then subtract the smallest value from the largest. For example, for the data set \(12, 14, 15, 16, 18, 19\):

\( \)Range\( = 19 - 12 = 7 \)

This range on its own only summarizes spread; it says nothing about where the data is centered. For that, you would look at measures like the mean or the ideas covered in center of a data set: mean, median, mode.

An outlier is a value that lies unusually far from the rest of the data. It might come from a measurement error, a rare event, or simply a genuinely extreme case. Outliers matter because they can pull statistics like the range or the mean far away from what most of the data actually looks like.

Suppose you add one more value to the earlier data set: \(12, 14, 15, 16, 18, 19, 45\). That \(45\) is far larger than every other value, so it looks like a candidate outlier.

Watch what happens to the range once that single value is included:

\( \)Range with outlier\( = 45 - 12 = 33 \)

\( \)Range without outlier\( = 19 - 12 = 7 \)

One extra value nearly quintupled the range, even though the rest of the data did not change at all. This is the main reason the range is called a "non-robust" measure of spread: a single extreme point can dominate it.

Looking at a number line can suggest an outlier, but a more precise test uses the interquartile range (\(IQR\)), which is the spread of the middle 50% of the data. For the sorted set \(12, 14, 15, 16, 18, 19, 45\):

The lower quartile is \(Q_1 = 14\) and the upper quartile is \(Q_3 = 19\), so \(IQR = Q_3 - Q_1 = 19 - 14 = 5\).

A commonly used boundary for a high outlier is:

\( \)Upper limit\( = Q_3 + 1.5 \times IQR = 19 + 1.5(5) = 26.5 \)

Since \(45\) is greater than \(26.5\), it is flagged as a true outlier rather than just a large but ordinary value. The same idea works on the low side using \(Q_1 - 1.5 \times IQR\). This rule is often shown visually with a box plot, where points beyond the "whiskers" are drawn as separate dots.

A small class scored these marks out of 50 on a quiz: \(30, 32, 33, 35, 36, 38, 8\).

Step 1: Order the data. \(8, 30, 32, 33, 35, 36, 38\)

Step 2: Find the range including every value. \(38 - 8 = 30\)

Step 3: Check for an outlier. The value \(8\) is far below the rest of the class, so it is a candidate outlier.

Step 4: Recalculate the range without it. \(38 - 30 = 8\)

Reporting a range of \(30\) would suggest the class scores were wildly spread out, when in reality one unusual score of \(8\) is doing all the work. Recognizing this lets you describe the data more honestly, for instance by reporting the range of the typical scores separately from the outlier.

Range and outliers come up constantly outside the classroom, from quality control on a factory line to spotting an unusually high expense in a budget. Because the range reacts so strongly to extreme values, it is usually reported alongside other summaries, such as the application of averages, so a data set is described fairly rather than by its most extreme point alone.

Related lessons

• Statistics Help