Anyone good with Statistics?

[mojo]

This might be a long shot as this is kind of a specific question, but there is a wide variety of people on this board.

When trying to calculate an overall standard deviation for a group of numbers, there is a specific formula that needs to used. You can not just take a weighted average of the individual standard deviations.

Does anyone know if this is also the case for semi-standard deviation?

[Vig]

This might be a long shot as this is kind of a specific question, but there is a wide variety of people on this board.

When trying to calculate an overall standard deviation for a group of numbers, there is a specific formula that needs to used. You can not just take a weighted average of the individual standard deviations.

Does anyone know if this is also the case for semi-standard deviation?

Is this what you're looking for?


Semi-standard deviation should work the same way, but only using numbers that fall below the mean.

[mojo]

Is this what you're looking for?


Semi-standard deviation should work the same way, but only using numbers that fall below the mean.

That formula gives me how to calculate standard deviation for 1 set of numbers. I'm trying to get a group number.

For standard deviation of a group (of two) of number you have to use this formula:
s2 group = wA2 s2A + (1 - wA)2 s2B + 2 wA wB rAB sA sB

Where all the "2" are actually squares except for the bolded one and s2 = group standard deviation. The more I look into it, the more I think I have to use this formula.

[VeryPrivate]

Sounds like you are describing the covariance.
http://en.wikipedia.org/wiki/Covariance

[mojo]

rAB sA sB

That's actually a formula for covariance ... which states:
correlation of A and B * variance of A * variance of B

[D-FENS]

nerds

10 chars

[mojo]

nerds

10 chars

Yep ... get that all the time.