Standard Deviation Calculator – Sample & Population, With Steps

Paste a list of numbers and choose sample or population. The calculator shows the standard deviation together with the variance, mean, standard error and coefficient of variation, and it works through every step in a table. A dot plot marks the mean and shades the area within one and two standard deviations, so you can see how spread out your data really is.

Standard Deviation Calculator

Separate values with commas, spaces or new lines. Paste a spreadsheet column straight in. For repeats, 7x3 means 7 three times.

Sample standard deviation–
Step by step

    A Worked Example, Step by Step

    Eight students score 4, 8, 6, 5, 3, 7, 9 and 6 on a 10-point quiz.

    1. Find the mean. The scores add up to 48, and 48 ÷ 8 = 6. (The Mean, Median, Mode Calculator covers this step in more depth.)

    2. Subtract the mean from each score and square the result.

    ScoreScore − 6Squared
    4−24
    824
    600
    5−11
    3−39
    711
    939
    600
    Total028

    The middle column always adds up to 0, because values above the mean balance those below it. Squaring stops them cancelling out. The total of the last column, 28, is the sum of squares.

    3. Divide to get the variance.

    • Population: 28 ÷ 8 = 3.5
    • Sample: 28 ÷ (8 − 1) = 4

    4. Take the square root.

    • Population standard deviation: √3.5 = 1.8708
    • Sample standard deviation: √4 = 2

    So a typical score sits about 2 points from the class average of 6.

    The standard deviation is in the same units as your data, points here. The variance is in squared units (points²), which is why people usually report the standard deviation instead: you can read it directly against the numbers you started with.

    Why the Sample Formula Divides by n − 1

    When your numbers are a sample, you don’t know the true mean of the whole group, so you measure distances from your sample’s own mean. Your sample’s mean is the one point that makes the sum of squares as small as possible for those values, so the sum of squares comes out a little smaller than it would around the true mean.

    You can check it with the quiz scores. Around their own mean of 6, the sum of squares is 28. Around any other center, such as 6.5, it’s bigger: 30. Dividing by n − 1 instead of n makes up for that shortfall. The adjustment has a name, Bessel’s correction.

    It matters most for small samples:

    Number of valuesSample variance is larger bySample SD is larger by
    525%11.8%
    1011.1%5.4%
    303.4%1.7%
    1001%0.5%

    With a few dozen values or more, the two answers are almost the same.

    Sample or Population?

    Ask one question: are these numbers everyone you care about, or a part you’re using to describe a bigger group?

    • Population: every score in the class you teach, the monthly sales for each of last year’s 12 months, all employees’ salaries at one company.
    • Sample: 200 customers surveyed to describe all customers, 30 bolts pulled from a production run, a week of readings used to estimate typical conditions.

    Most real data is a sample, which is why the calculator starts on Sample. Use Population when the list really is the whole group and you only want to describe it.

    What the 68–95–99.7 Rule Says

    For data shaped like a bell curve, about 68% of values fall within one standard deviation of the mean, about 95% within two, and about 99.7% within three. You’ll also see the same idea used to say how unusual a single value is:

    Distance above the meanShare of values below itShare within that distance on both sides
    1 SD84.1%68.3%
    1.645 SD95%90%
    2 SD97.7%95.4%
    3 SD99.87%99.7%

    A value 1.645 standard deviations above the mean is in the top 5%, not the top 10%. The 90% figure only applies when you count both tails together.

    These percentages only hold for bell-shaped data. That’s why the calculator counts your actual values instead of assuming them. In the quiz example, 6 of the 8 scores (75%) are within one standard deviation of the mean, and all 8 are within two.

    How Far Is One Value From the Mean?

    Divide its distance from the mean by the standard deviation. This is a z-score. In the quiz, a score of 9 is 3 points above the mean of 6, and 3 ÷ 2 = 1.5, so it sits 1.5 standard deviations above average. A score of 3 has a z-score of −1.5. Z-scores let you compare results from tests with different averages and spreads. The z-score calculator turns any score into a z-score and a percentile.

    One Extreme Value Can Inflate It

    Because the deviations are squared, values far from the mean have an outsized effect. Change the 9 in the quiz scores to 29, say through a typing error, and the sample standard deviation jumps from 2 to about 8.43. The median stays at 6.

    If the standard deviation seems too large, look at the dot plot for a value sitting far from the rest. It might be a mistake or a genuine outlier, and either way it’s worth knowing about before you report the number.

    Standard Deviation vs Standard Error

    The two are easy to confuse:

    • Standard deviation describes how spread out the individual values are.
    • Standard error of the mean describes how precisely your sample’s mean estimates the true mean. It equals the standard deviation divided by √n.

    For the quiz, the standard error is 2 ÷ √8 ≈ 0.71. Collect more data and the standard error shrinks, because the mean becomes more reliable. The standard deviation doesn’t shrink, because individual students don’t become more alike just because you measured more of them.

    Comparing Spread Across Different Scales

    A standard deviation of 2 points means a lot on a 10-point quiz and very little on a 100-point exam. To compare the two, divide the standard deviation by the mean. This is the coefficient of variation (CV):

    • Quiz: 2 ÷ 6 = 33.3%
    • Exam with a mean of 70 and a standard deviation of 8: 8 ÷ 70 = 11.4%

    The quiz scores are far more spread out relative to their average, even though the exam’s standard deviation is four times larger. Whether a spread is too high depends on what you measure; what is a good standard deviation walks through four checks. The CV only makes sense for data that can’t go below zero, such as scores, weights or prices. For data with negative values, like temperatures in Celsius, the calculator leaves it out.