Re: rusty sigmas

posted by [identity profile] purplecthulhu.livejournal.com at 11:25am on 15/03/2009
I'm not quite sure I understand you here. We don't know what the expected number of results is since we don't know the parent distribution - it's what we're trying to determine - so the best estimate we have for that is the results obtained. Hence the sigma on the parent mean is just the sqrt of the number of such events.

To me more rigorous I should be adding the errors in quadrature which does this:

9.3 children in test (sd = 3) vs. 2.5 (sd = 1.6) in control were much improved. Difference is 6.8. Adding errors in quadrature is sqrt(3^2 + 1.6^2) = 3.4 giving a final significance of 2 sigma (happens one time in 20 at random).

24.8 (sd=4.98) vs. 11.78 (sd = 3.4) improved. Different is 13.02 error added in quadrature is 6.04 so the final result is 2.15 sigma, so still insignificant.

Numbers change a bit (and I was being a bit sloppy I admit) but the conclusion remains the same: no result.

[parent entry] [permalink]

Reply

This account has disabled anonymous posting.
If you don't have an account you can create one now.
HTML doesn't work in the subject.
More info about formatting
 
Notice: This account is set to log the IP addresses of everyone who comments.
Links will be displayed as unclickable URLs to help prevent spam.

April

SunMonTueWedThuFriSat
    1
 
 2
 
 3
 
 4
 
 5
 
6
 
 7
 
 8
 
 9
 
 10
 
 11
 
 12
 
13
 
 14
1
 15
 
 16
 
 17
 
 18
 
 19
 
20
 
 21
 
 22
 
 23
 
 24
 
 25
 
 26
 
27
 
 28
 
 29
 
 30