PDA

View Full Version : a question

stormhold
August 28th 2009, 04:35 PM
I am in the lognormal distribution chapter and saw a problem about estimating the stock's annualized return. the book says
http://farm3.static.flickr.com/2622/3865939260_2109d578c7.jpg

I know VAR(X)=E(X^2)-(E(X))^2. But on the book, why there's a n/(n-1)?

Thanks.

iromio
August 28th 2009, 04:41 PM
I am in the lognormal distribution chapter and saw a problem about estimating the stock's annualized return. the book says
http://farm3.static.flickr.com/2622/3865939260_2109d578c7.jpg

I know VAR(X)=E(X^2)-(E(X))^2. But on the book, why there's a n/(n-1)?

Thanks.

i too stuck on this....

NoMoreExams
August 28th 2009, 04:48 PM
Think back to stats 101, if you divide by n, you get population variance. Obviously you don't have "all" the data, you have a sample so you need to convert it to to sample variance, hence the n/(n-1) factor.

stormhold
August 28th 2009, 04:57 PM
Think back to stats 101, if you divide by n, you get population variance. Obviously you don't have "all" the data, you have a sample so you need to convert it to to sample variance, hence the n/(n-1) factor.

Can you give me a link that further explains this?

NoMoreExams
August 28th 2009, 05:04 PM
Does this help: http://en.wikipedia.org/wiki/Variance#Population_variance_and_sample_variance

rahim
August 29th 2009, 04:15 AM
n/n-1 is multiplied to make the variance unbiased.
you can find this in any mathematical statistics book