If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

### Course: Statistics and probability>Unit 3

Lesson 6: More on standard deviation

# Unbiased estimate of population variance

A CS program to help build intuition
Make a spinoff here!

## Want to join the conversation?

• The title of this page is, "Unbiased estimate of population variance", but doesn't this program measure the biased estimate?
• as the title says, it is about "estimating" the unbiased value using biased value. with sample sizes from 2 to 10, it shows a relation of (n-1)/n between the two, resulting in the division with the "n-1". for a sample size of 2 this is 1/2, and of 3 gives 2/3 and so on. it becomes "unbiased = biased *n/(n-1)" or simply the equation with "n-1" as the denominator.
• So what I'm getting is that the n-1 unbiased formula describes the phenomenon where the sample variance estimate is closer to population variance using n-1 than n. However, I don't understand why this phenomenon occurs? Why does using n to divide cause an underestimation in the first place?
• The underestimation arises from the fact that using n (the sample size) as the divisor doesn't fully account for the variability present in the sample (because n<N).
• In Population, N=463 but sample Size Vs variance has samples in thousands ?? How come it is possible?
• think about how many groups (=samples) of 2, 3, 4...or 10 you could build out of 463! Lets say it would be 463 people and you build groups of 2. How many different groups could you build? Combinatorics say 463C2, which is 463*462/2 which is 106,953 possibilities. But then you can also build groups of 3, of 4... and so on. There is a lot of possibilities to build samples!
• how do i find the variance of classified data?
• how do i find the variance of classified data?
• yeap this is biased one