Variance Calculator

When the Average Is Not Enough

Priya runs quality control at a bottling plant. Her line fills 500 ml bottles, and the target mean is exactly 500 ml. Last week the average fill came in at 500.1 ml, which looked perfect. But the plant had received three customer complaints about underfilled bottles. The mean hid the problem. Some bottles held 488 ml while others held 512 ml. They averaged out to roughly 500 ml, yet the spread was wide enough to cause real issues. Variance is the number that captures that spread. It tells you how far the individual values sit from the mean, on average, in squared units. This variance calculator gives you that number in seconds, along with the full step-by-step breakdown so you can see exactly where it comes from.

What This Calculator Does and Why It Matters

This tool computes variance from any list of numbers. You type in your data, pick a mode, and the calculator returns the variance, the standard deviation, the mean, the count, and the sum of squared deviations. It also shows a row-by-row breakdown of each value, its deviation from the mean, and its squared deviation. That breakdown is the part most textbooks skip. Seeing the arithmetic worked out term by term is what turns a formula into something you actually understand.

Variance is the companion to standard deviation. In fact, standard deviation is just the square root of variance. People often reach for standard deviation first because it shares the same units as the original data. But variance is the quantity that shows up in the underlying math. It appears in formulas for confidence intervals, hypothesis tests, regression, and ANOVA. If you want the standard deviation directly, use our Standard Deviation Calculator. For a broader set of summary statistics, the Statistics Calculator covers mean, median, range, and more in one pass.

• Population mode: divides the sum of squared deviations by n. Use this when your data is the entire group you care about.
• Sample mode: divides by n - 1. Use this when your data is a sample drawn from a larger population and you want an unbiased estimate of that population's variance.
• Step-by-step breakdown: shows every value, its deviation from the mean, and its squared deviation, then the final division.

How the Calculation Works

Variance measures the average squared distance between each data point and the mean. The formula differs depending on whether you are working with a full population or a sample.

For a population, the variance is written as σ²:

For a sample, the variance is written as s²:

Here x_i is each individual value, μ is the population mean, x̄ is the sample mean, and n is the number of values. The only difference between the two formulas is the denominator. The population formula divides by n. The sample formula divides by n - 1.

That n - 1 is called Bessel's correction. When you compute variance from a sample, you use the sample mean as your reference point. The sample mean is always closer to the sample data than the true population mean is, because the sample mean is built from that same data. This makes the squared deviations a bit too small on average, which biases the result downward. Dividing by n - 1 instead of n corrects for that bias. The correction matters most when your sample is small. With 200 values the difference between n and n - 1 is tiny. With 5 values it is large.

Once you have the variance, the standard deviation is just its square root. That is why the two measures always tell the same story in different units. You can explore that relationship in depth with the Standard Deviation Calculator.

How to Use the Calculator

1. Choose a mode at the top: Population or Sample. Pick Population if your data is the entire group. Pick Sample if your data is drawn from a larger population.
2. Type your numbers into the text box. Separate them with commas or spaces. The default example is 2, 4, 4, 4, 5, 5, 7, 9.
3. The calculator detects how many values you entered and shows that count below the input.
4. Read the primary result card for the variance in your selected mode. The summary card shows the other variance, the standard deviation, the mean, the count, and the sum of squared deviations.
5. Scroll to the step-by-step breakdown to see each value, its deviation from the mean, and its squared deviation, followed by the final division.
6. Use the Copy Result button to save the output for your notes or report.

Example Calculations

Example 1: Priya's bottling line. She samples 6 bottles and measures 498, 502, 495, 505, 500, 504 ml. The mean is (498 + 502 + 495 + 505 + 500 + 504) / 6 = 3004 / 6 = 500.67 ml. The deviations are -2.67, 1.33, -5.67, 4.33, -0.67, and 3.33. Squaring those gives 7.11, 1.78, 32.11, 18.78, 0.44, and 11.11. The sum of squared deviations is 71.33. Since this is a sample from the production run, she uses sample variance: 71.33 / (6 - 1) = 71.33 / 5 = 14.27. The sample standard deviation is the square root, about 3.78 ml. That tells her the fill amounts typically deviate by roughly 3.8 ml from the mean, which is within the plant's tolerance of 5 ml.

Example 2: Marcus grades his class. A teacher named Marcus has 8 students and their exam scores are 72, 85, 90, 68, 77, 82, 95, 79. He wants the population variance because these 8 students are the entire class, not a sample. The mean is (72 + 85 + 90 + 68 + 77 + 82 + 95 + 79) / 8 = 648 / 8 = 81. The deviations from 81 are -9, 4, 9, -13, -4, 1, 14, and -2. Squaring those gives 81, 16, 81, 169, 16, 1, 196, and 4. The sum is 564. Population variance is 564 / 8 = 70.5. The population standard deviation is about 8.4 points. Marcus can see that the scores spread fairly widely around the 81 average, which tells him the class was not clustered tightly around the mean.

Real-World Scenarios

Investment risk. Portfolio managers use variance to measure the volatility of returns. A stock whose monthly returns swing between -8% and +12% has a higher variance than one that stays between -1% and +3%, even if both average the same return. Investors often prefer lower variance for the same expected return, because the ride is smoother. Variance is the backbone of modern portfolio theory, where it is used to balance risk across assets. When you hear a fund described as "high risk," variance (or its square root, standard deviation) is usually the number behind that label.

Manufacturing tolerance. Engineers set tolerance limits on machined parts. A shaft might need a diameter of 25 mm with a tolerance of plus or minus 0.05 mm. By measuring a batch of shafts and computing the variance, the engineering team can tell whether the production process is staying within spec. A rising variance signals that something in the process is drifting, whether it is tool wear, material variation, or operator handling. Catching that drift early prevents a whole batch from failing inspection.

Exam score spread. Teachers and education researchers use variance to understand how widely scores distribute. A class with a mean of 75 and a low variance means most students scored near 75. The same mean with a high variance means some students scored very high and others very low. That difference changes how a teacher responds. A tight cluster might call for moving on to new material. A wide spread might call for review sessions or differentiated instruction. For the full picture of a data set's center and spread together, the Mean Median Mode Range Calculator is a good companion tool.

Common Mistakes to Avoid

• Confusing population and sample variance. If your data is a sample, use n - 1. If it is the entire population, use n. Using the wrong one gives a biased estimate. The calculator lets you switch modes with one click so you can compare both values side by side.
• Forgetting to square the deviations. The deviations themselves sum to zero, because the positive and negative distances from the mean cancel out. Squaring each deviation before summing is what makes the math work. If you skip the squaring step, you get zero every time.
• Dividing by n when you should divide by n - 1. This is the most common error in introductory statistics. It underestimates the population variance when working from a sample. Bessel's correction exists precisely to fix this. The difference shrinks as your sample grows, but for small samples it is substantial.
• Mixing up variance and standard deviation. Variance is in squared units. Standard deviation is in the original units. A variance of 14.27 ml² does not mean the bottles vary by 14.27 ml. They vary by the square root of that, about 3.78 ml. Always take the square root before interpreting the spread in real-world units.

Limitations

Variance is sensitive to outliers. Because each deviation is squared, a single extreme value can inflate the result dramatically. If one bottle in Priya's sample held 600 ml instead of 504 ml, the variance would jump from 14.27 to over 1,800. That is why statisticians often report the median and interquartile range alongside variance when outliers are present. Variance is also expressed in squared units, which can be hard to interpret on its own. A variance of 70.5 points² on an exam does not translate intuitively the way a standard deviation of 8.4 points does. Finally, variance tells you the size of the spread but not its direction. Two data sets with the same variance can have completely different shapes, one symmetric and one skewed. Variance is a summary of spread, not a full description of the distribution.

Authoritative Research and Resources

The NIST/SEMATECH e-Handbook of Statistical Methods covers variance, standard deviation, and their role in exploratory data analysis. It is a practical reference used by engineers and scientists, with worked examples and clear definitions.

For a guided walkthrough of variance and why Bessel's correction matters, Khan Academy's lesson on population and sample variance explains the two formulas, the intuition behind n - 1, and practice problems with step-by-step solutions.

The OECD Glossary of Statistical Terms provides a formal definition of variance used across official statistics. It is a good reference when you need the precise terminology for a report or academic paper.