The variance is the average of the squared deviations from the mean; it equals the standard deviation squared.
σ2=n∑(x−μ)2
Quick Answer
The variance is the average of the squared deviations from the mean; it equals the standard deviation squared. Use it when spreads have to be combined or compared inside a calculation: for independent random quantities, variances add while standard deviations do not. Compute each deviation from the mean, square it, and average the squares, dividing by n for a population or n−1 for a sample. Its units are squared, so take the square root when you describe the spread to a person.
Why This Matters
Variance is the form of spread that does arithmetic well. If two independent steps of a process have SDs of 3 and 4 minutes, the total time has variance 9+16=25, so its SD is 5 minutes, not 7. Without variance you would add standard deviations and overstate the combined uncertainty.
Intuitive Explanation
Take the data 2, 4, 6, 12. The mean is 6, and the deviations are −4,−2,0,6. Square each to get 16, 4, 0, 36, and average: 456=14. That 14 is the variance. The value 12, six above the mean, supplies 36 of the 56 on its own; squaring makes a value twice as far away count four times as much (compare the 4 and the 16).
Squaring also squares the units. If the data are in dollars, the variance is in square dollars, which has no everyday meaning. A test-score variance of 36 does not mean scores spread over 36 points; it means the standard deviation is 36=6 points. That is why people quote the standard deviation and keep the variance for formulas.
The payoff for squared units is that variances of independent quantities add. One fair die has variance about 2.92; the total of two dice has variance about 5.83, exactly double. The standard deviations, about 1.71 and 2.42, do not add: 1.71+1.71=3.42, not 2.42. Two things look like variance but are not: the average of the unsquared deviations, which is always 0, and the average of their absolute values, which is the mean absolute deviation.
Think of it like this.Draw a square on each deviation: a deviation of 3 becomes a 3-by-3 square with area 9. The variance is the average area of those squares. Areas add cleanly, which is why variances combine, but an area is not a length, so you take a square root to get back to a distance.
Core idea.Variance is the average squared distance from the mean — the spread measured in squared units.
When to Use
Use the variance when a problem asks for the average squared deviation, when you add or subtract independent random quantities (total time of two steps, the sum of two dice), or when a formula is written in terms of σ2 or s2. The word variance itself is the plainest signal. If the question asks how spread out the data is in its own units, give the standard deviation, the square root of the variance.
Tip.Ask: Am I averaging the squared distances from the mean (and not taking the square root)?
How to Recognize It
Before reaching for the variance, decide whether you need a spread to report to a person or a spread to calculate with, since only the second is worth keeping in squared units.
1.
Are you combining the spreads of two or more independent quantities, such as the total of two dice or two steps in a process?
If yes, add the variances, because standard deviations do not add. If you only need to describe one data set's spread, the standard deviation, in the data's units, is usually the better answer.
2.
Does the formula or the question use σ2 or s2?
If yes, it is asking for the variance, the number before the square root. If it uses σ or s, or asks how far values typically sit from the mean, take the root and give the standard deviation.
3.
Will someone read the answer as a size they can picture, like minutes or dollars?
If yes, a variance in square minutes or square dollars is hard to interpret, so convert it to the standard deviation. If the number feeds another calculation, keep the variance.
4.
Must every value count toward the spread, not just the largest and smallest?
If yes, the variance uses every value's distance from the mean. If a rough end-to-end width is enough, the range, maximum minus minimum, is quicker but ignores everything in between.
Variance compared with Standard deviation, Mean, and Range
Variance, standard deviation, mean and range appear in the same calculations, so one is easily passed off as another. Ask whether the number is a center or a spread, and if a spread, whether it is in squared units (variance) or the data's units (standard deviation, range).
Variance compared with Standard deviation, Mean, and Range
Type
Meaning
Key test
Formula
Example
Variance
Use this when you want a spread measure in squared units that adds and combines cleanly in formulas.
Am I averaging the squared distances from the mean (and not taking the square root)?
σ2=n∑(x−μ)2
Find the population variance of 1,3,5,7 (mean =4).
Standard deviation
The square root of variance — spread in the data's real units.
Use when reporting spread to people or comparing to actual values.
σ=σ2
SD 5 means variance 25
Mean
The center the deviations are measured from, not a spread.
Use when you want the balance point, which variance needs as input.
n∑x
Average before computing deviations
Range
Spread from extremes only; ignores how every value scatters.
Use for a quick spread when you do not need a squared, additive measure.
max−min
Quick width of a data set
Variance
Meaning
Use this when you want a spread measure in squared units that adds and combines cleanly in formulas.
Key test
Am I averaging the squared distances from the mean (and not taking the square root)?
Formula
σ2=n∑(x−μ)2
Example
Find the population variance of 1,3,5,7 (mean =4).
Standard deviation
Meaning
The square root of variance — spread in the data's real units.
Key test
Use when reporting spread to people or comparing to actual values.
Formula
σ=σ2
Example
SD 5 means variance 25
Mean
Meaning
The center the deviations are measured from, not a spread.
Key test
Use when you want the balance point, which variance needs as input.
Formula
n∑x
Example
Average before computing deviations
Range
Meaning
Spread from extremes only; ignores how every value scatters.
Key test
Use for a quick spread when you do not need a squared, additive measure.
How to read it:σ2 for population variance, s2 for sample variance
Worked Examples
Example 1 — Variance of four values
Easy
Find the population variance of 1,3,5,7 (mean =4).
Show the steps and the answer
Find the deviations from 4: −3,−1,1,3.
They add to 0, which confirms the mean and shows why raw deviations cannot simply be averaged.
Square them: 9,1,1,9.
Squaring makes every term positive and gives the outer values, 1 and 7, more weight.
Add and divide by n=4: 49+1+1+9=420=5.
The problem asks for the population variance, so the divisor is n.
Stop here, without a square root.
The root, 5≈2.24, would be the standard deviation, a different number.
Answer:5
Takeaway:Variance is the average squared deviation: the standard deviation before its final square root.
Example 2 — They want a reportable spread
Standard
Test-score variance is 36 and a parent asks 'how spread out are the scores?'
Show the steps and the answer
Identify the units: a variance of 36 is in squared points.
A parent cannot picture square points, so 36 is not a spread they can use.
Convert: 36=6 points.
The standard deviation is the square root of the variance and is measured in points, like the scores.
Explain it: a typical score is about 6 points above or below the class mean.
Quoting 36 would suggest a spread six times larger than the real one.
Answer:36=6 points
Takeaway:Keep the variance for calculations and convert to the standard deviation, with one square root, when you explain spread.
Example 3 — Find the mistake: deviations that cancel
Application
The data are 3, 5 and 10, with mean 6. Kai writes: "The deviations are −3, −1 and 4. Their average is 0, so the variance is 0 and the data has no spread." Find the mistake and correct it.
Show the steps and the answer
Check Kai's deviations: 3−6=−3, 5−6=−1, 10−6=4, and −3−1+4=0.
Deviations from the mean add to zero for every data set, so their average says nothing about spread.
Test the conclusion: 3, 5 and 10 are clearly different values.
A variance of 0 happens only when every value equals the mean.
Square the deviations: 9,1,16, total 26.
Squaring stops the negative deviations from cancelling the positive ones.
Average over n=3: 326≈8.67.
This is the population variance of the three values.
Answer:Variance =326≈8.67
Takeaway:Square the deviations before averaging; unsquared deviations always cancel to zero.
Where You'll See It
Planning a two-part commute
A bus ride averages 20 minutes with SD 4, and the walk after it averages 10 minutes with SD 3, independently. Variances add, 16+9=25, so the whole trip has SD 25=5 minutes, not 4+3=7. Planning around 7 would leave more slack than you need.
Why spreading investments lowers risk
Suppose two independent stocks each have return variance v. Putting half your money in each gives a portfolio variance of 4v+4v=2v, half the risk of owning one stock. This variance calculation is the basic math behind diversification.
Common Mistakes
The slip:Taking the square root and calling it variance
Instead:that result is the standard deviation; variance is before the root.
The slip:Averaging the deviations without squaring
Instead:raw deviations sum to zero, so square them first.
The slip:Mixing up the divisor
Instead:population variance divides by n, sample variance (s2) divides by n−1.
Mini Practice
Try these on your own. Tap Reveal when you want to check.
1.
Find the population variance of 3,5,7,9,11. The mean is 7.
Hint:Square each distance from the mean, then average the squares.
2.
Find the population variance of 2,3,7,8.
Hint:Find the mean first; it is a whole number here.
3.
A sample of three measurements is 4,7,10 kg. Find the sample variance s2.
Hint:A sample variance divides by one less than the number of values.
4.
A data set has variance 5. Every value is doubled. What is the new variance?
Hint:Doubling each value doubles each deviation. What happens to a squared deviation?
5.
Two independent tasks have times with standard deviations 3 min and 4 min. To find the spread of the total time, should you add the standard deviations or the variances?
Hint:Only one of these two spread measures adds directly for independent quantities.
6.
For 1,4,7 (mean 4), a student writes: "The deviations are −3,0,3. Their average is 0, so the variance is 0." Correct the work.
Hint:What must happen to each deviation before averaging?
Want the full set?
50 practice questions for this concept — free to try, every one with a complete worked
solution showing the why, not just the answer.
Why do variances add when standard deviations do not?
For independent quantities, spreads combine like the two shorter sides of a right triangle. If one step's standard deviation is 3 minutes and another's is 4, the total's standard deviation is 32+42=5 minutes, not 7. Squaring turns that into plain addition: 9+16=25. In everyday terms, a slow first step is often offset by a quick second one, so the total varies less than the two spreads stacked end to end.
What happens to the variance if I double every value?
It is multiplied by 4. Every deviation from the mean doubles, and squaring a doubled deviation gives four times as much. Adding the same number to every value changes nothing, because the mean moves too and the deviations stay the same. Changing units has the same squared effect: converting lengths from meters to centimeters multiplies the variance by 1002=10,000, while the standard deviation grows only by 100.
Can the variance be smaller than the standard deviation?
Yes, whenever the standard deviation is less than 1. A standard deviation of 0.5 gives a variance of 0.52=0.25, because squaring a number between 0 and 1 makes it smaller. Only when the standard deviation is above 1 is the variance the larger of the two. Comparing them is not very meaningful anyway, since they are in different units, such as centimeters and square centimeters.
Is there a quicker way to calculate a variance by hand?
Yes: average the squares, then subtract the square of the mean, σ2=x2−xˉ2. For 1, 2 and 3, the mean is 2 and the mean of the squares is 31+4+9=314, so σ2=314−4=32, the same as squaring the deviations −1,0,1. It saves subtracting the mean from every value, but with large numbers and rounding it can lose accuracy.
You need the Mean, since deviations are measured from it, and Standard Deviation, the square root of the variance. Variance adds the algebra of spread: squared deviations, and the rule that independent variances add. It feeds back into Standard Deviation, which you report once the variance is known, and it is the building block for combining random variables and for regression later on.