Range vs standard deviation

By Jude Wallis · Published

The range is the maximum minus the minimum, so it depends on two values only. The standard deviation measures the typical distance of all the values from the mean. Two data sets can share a range of 40 and have standard deviations of 12.7 and 19.8. Both measures are nonresistant to outliers.

AP Statistics: Unit 1 (topics 1.7 Summary Statistics for One Quantitative Variable). The range and the standard deviation are measures of variability in topic 1.7, Summary Statistics for One Quantitative Variable, in Unit 1 of the Fall 2026 AP Statistics course, which describes both as nonresistant because outliers can affect their values greatly. The interquartile range is the resistant measure in the same topic.

Range vs standard deviation: the short answer

Both are measures of variability for a quantitative variable, and both are reported in the units of the data. They differ in how much of the data they use. The range is one subtraction, the largest value minus the smallest, so the values between the two extremes have no effect on it. The standard deviation is built from the distance of every value from the mean, so it reflects how the whole data set is arranged.

That difference shows up whenever two data sets have the same extremes and different interiors. The range reports them as equally variable. The standard deviation does not. Both measures share one weakness: a single extreme value changes them substantially, so neither is resistant.

What the range measures

The range is a single number:

range=maximum−minimum\text{range} = \text{maximum} - \text{minimum}

It gives the total width of the interval the data occupy. It is quick to compute and easy to explain, which is why it appears in weather reports and quality checks.

Its limits come from using only two observations. It ignores how the other values are distributed between the extremes. It is set entirely by the two most unusual values in the data, which are the ones most likely to be errors. And it tends to grow with sample size: adding an observation can widen the range or leave it alone, and can never shrink it, so a sample of 200 will usually show a larger range than a sample of 20 from the same population. That last property is examined in range vs IQR.

What the standard deviation measures

The sample standard deviation, written ss, is

s=∑(xi−xˉ)2n−1s = \sqrt{\frac{\sum (x_i - \bar{x})^2}{n - 1}}

where xˉ\bar{x} ("x-bar") is the sample mean and nn is the number of values. Each value's deviation from the mean is squared, the squares are added and divided by n−1n - 1, and the square root returns the result to the original units. A step-by-step version is in standard deviation by hand.

The standard deviation is read as the typical distance of the values from their mean. Because every observation contributes, moving any value changes it. It is the measure of variability that pairs with the mean, and it is the one used in zz-scores, the normal distribution, and the standard errors of inference.

Squaring gives large deviations extra weight, so the standard deviation is sensitive to outliers. Unlike the range, it has no built-in tendency to grow with sample size: adding an observation near the mean lowers it, and adding one far from the mean raises it.

The differences side by side

FeatureRangeStandard deviation
FormulaMaximum minus minimum∑(xi−xˉ)2/(n−1)\sqrt{\sum (x_i - \bar{x})^2 / (n - 1)}
Values usedTwoAll nn
What it describesTotal width of the dataTypical distance from the mean
UnitsUnits of the dataUnits of the data
Resistant to outliersNoNo
Effect of adding an observationStays the same or growsCan go up or down
Sensitive to values between the extremesNoYes
Used in later calculationsRarelyzz-scores, normal models, standard errors

For any data set with at least two different values, the standard deviation is smaller than the range. A computed standard deviation larger than the range signals an arithmetic error.

When to use which

Use the standard deviation as the working measure of variability whenever the mean is the measure of center, which is the usual case for roughly symmetric data without outliers. It is also the only one of the two that later topics build on.

Use the range as a quick supporting fact. Together with the minimum and maximum it tells a reader the extent of the data, and it is a useful check on data entry: an impossible range, such as 918 points on a 100-point test, reveals a typing error at once. It is a weak basis for comparing data sets of different sizes, for the sample-size reason given above.

When the data are skewed or contain outliers, neither is the best choice. The interquartile range is resistant, and standard deviation vs IQR covers that decision. You can compute all three at once with the standard deviation calculator and the five-number summary calculator.

The classic mix-up and how to avoid it

The first mistake is writing the range as an interval. "The range is 50 to 90" reports the minimum and the maximum. The range is their difference, 40.

The second mistake is assuming equal ranges mean equal variability. The first worked example has two data sets with minimum 50, maximum 90, and mean 70. In one, the middle values sit close to the mean; in the other, they sit near the extremes. The ranges match at 40 and the standard deviations are 12.7 and 19.8.

The third mistake is believing the standard deviation is the outlier-proof alternative to the range. It is not resistant. In the second worked example, changing one value from 8 to 28 multiplies the range by 6 and the standard deviation by about 6.4. If an outlier is the concern, the resistant measure is the IQR.

Same range, different standard deviations

Class A scores are 50, 68, 70, 70, 72, 90. Class B scores are 50, 52, 54, 86, 88, 90. Find the range and the sample standard deviation of each class.

  1. Ranges: class A is 90−50=4090 - 50 = 40 and class B is 90−50=4090 - 50 = 40.

  2. Means: class A sums to 50+68+70+70+72+90=42050 + 68 + 70 + 70 + 72 + 90 = 420, so xˉ=4206=70\bar{x} = \frac{420}{6} = 70. Class B sums to 50+52+54+86+88+90=42050 + 52 + 54 + 86 + 88 + 90 = 420, so xˉ=70\bar{x} = 70 as well.

  3. Class A deviations from 70: −20,−2,0,0,2,20-20, -2, 0, 0, 2, 20. Squares: 400,4,0,0,4,400400, 4, 0, 0, 4, 400, which sum to 808.

  4. Class A standard deviation: s=8085=161.6≈12.71s = \sqrt{\frac{808}{5}} = \sqrt{161.6} \approx 12.71.

  5. Class B deviations from 70: −20,−18,−16,16,18,20-20, -18, -16, 16, 18, 20. Squares: 400,324,256,256,324,400400, 324, 256, 256, 324, 400, which sum to 1960.

  6. Class B standard deviation: s=19605=392≈19.80s = \sqrt{\frac{1960}{5}} = \sqrt{392} \approx 19.80.

  7. Compare: both classes have range 40 and mean 70. Class B's standard deviation is larger because four of its six scores are 16 or 18 points from the mean, while class A's middle four scores are within 2 points of it.

Both classes have a range of 40 points. The standard deviation is about 12.71 points for class A and about 19.80 points for class B. The range sees only the 50 and the 90, which the classes share. The standard deviation uses all six scores, so it registers that class B's scores sit farther from the mean.

One extreme value changes both

Five values are 4, 5, 6, 7, 8. Find the range and the sample standard deviation. Then change the 8 to 28 and find both again.

  1. Original range: 8−4=48 - 4 = 4.

  2. Original mean: 4+5+6+7+85=305=6\frac{4 + 5 + 6 + 7 + 8}{5} = \frac{30}{5} = 6. Deviations: −2,−1,0,1,2-2, -1, 0, 1, 2. Squares sum to 4+1+0+1+4=104 + 1 + 0 + 1 + 4 = 10.

  3. Original standard deviation: s=104=2.5≈1.58s = \sqrt{\frac{10}{4}} = \sqrt{2.5} \approx 1.58.

  4. New data: 4, 5, 6, 7, 28. New range: 28−4=2428 - 4 = 24.

  5. New mean: 4+5+6+7+285=505=10\frac{4 + 5 + 6 + 7 + 28}{5} = \frac{50}{5} = 10. Deviations: −6,−5,−4,−3,18-6, -5, -4, -3, 18. Squares sum to 36+25+16+9+324=41036 + 25 + 16 + 9 + 324 = 410.

  6. New standard deviation: s=4104=102.5≈10.12s = \sqrt{\frac{410}{4}} = \sqrt{102.5} \approx 10.12.

  7. Compare: the range went from 4 to 24, six times as large. The standard deviation went from 1.58 to 10.12, about 6.4 times as large. The median is 6 in both versions.

The range rises from 4 to 24 and the standard deviation rises from about 1.58 to about 10.12. One changed value multiplied both measures by a factor of about 6, while the median stayed at 6. Neither the range nor the standard deviation is resistant to an extreme value.

Frequently asked questions

Can the standard deviation be larger than the range?

No. A sample standard deviation is at most the range divided by 2\sqrt{2}, about 71% of the range, and it reaches that ceiling only when the sample has just two values. With more data it is usually much smaller than the range.

Which is more affected by outliers, the range or the standard deviation?

Both are strongly affected, and neither is resistant. The range is set directly by the most extreme values. The standard deviation squares each deviation, so one far-off value can dominate the sum. For a resistant measure of variability, use the interquartile range.

If two data sets have the same range, do they have the same standard deviation?

Not in general. The range fixes only the distance between the two extremes. The scores 50, 68, 70, 70, 72, 90 and 50, 52, 54, 86, 88, 90 both have range 40, with standard deviations of about 12.71 and 19.80.