combined variance

[Monica Muller]

Hi All,

I need to calculate combined variance of several samples. I only have mean, standard deviation, and sample size of the samples. Is there anyway to do that in Stata with this information?

Thanks!

[Carlo Lazzaro]

Monica:
can you please provide the list with an example of what you're arfter? Thanks.

[Monica Muller]

Sure! Here is a small part of my data:

each row shows mean, sample size(n), and standard error of one sample for two different variables (ab_A3IM and Ab_A4IM)

I need to find the overall mean and standard deviation for the population that includes these samples. (N=152)

I found a formula (the 5 steps at the end of the page) for it but I don't know how to apply it on my data:

clear
input int id     int   ab_meanA3IM     int   ab_nA3IM      int  ab_seA3IM   int     ab_meanA4IM        int ab_nA4IM      int  ab_seA4IM
id        ab_meanA3IM        ab_nA3IM        ab_seA3IM        ab_meanA4IM        ab_nA4IM        ab_seA4IM
1        4.38        8        0.18        4.12        8        0.13
2        4        6        0        4        6        0
3        4        8        0.19        3.62        8        0.18
5        4.38        9        0.18        3.88        9        0.13
7        4        12        0.19        3.88        12        0.13
8        4.12        8        0.13        3.75        8        0.16
9        4.5        10        0.19        3.75        10        0.16
10        4        9        0.19        3.5        9        0.19
11        4.12        8        0.13        3.88        8        0.13
13        4        7        0        4.12        7        0.13
14        4.12        8        0.13        4        8        0
15        3.88        9        0.13        4        9        0
16        4.12        10        0.13        3.88        10        0.13
17        4.12        8        0.13        4        8        0
19        3.88        8        0.13        3.88        8        0.13
20        3.88        8        0.13        3.75        8        0.16
22        4        8        0        4        8        0
23        3.88        8        0.3        3.25        8        0.16
end

I really appreciate your help

[Marcos Almeida]

I'm not sure if I got your point quite right, but, to start, I gather you could play with - egen - with rowmean and rowsd. For example:

. egen newvar1 = rowmean(var1 var2 var3)
. egen newvar2 = rowsd(var1 var2 var3)

Then, with - generate - you could convert the sd into variance, just by squaring the sd.


Hopefully that helps.

[Biswa bhusan]

Dear Marcos, I have to find out the combined variance of three variables. However, i am unable to understand above data set and explanation. Using following data set:
sysuse auto.dta

How to find out combined variance between price, mpg, and displacement using Stata?