Re: st: Sample size calculation

[Ron�n Conroy <rconroy@rcsi.ie>]

On 26 DF�mh 2005, at 20:34, <slaowattana@pol.net>  
<slaowattana@pol.net> wrote:


> Thanks. We plan to do the experiment in humans and they will be  
> treated
> with Rebif. The expected size of difference is 2 SD (to be clinically
> significant) but my understanding is that to be able to calculate the
> sample size (alpha = 0.05, beta = 0.20), we have to know estimated  
> mean
> and SD? Is this not true?
No, no. Sample size calculations are done on the difference between  
the two groups expressed in standard deviations. The actual size of  
the SD is irrelevant.

. sampsi 0 2, sd(1)

Estimated sample size for two-sample comparison of means

Test Ho: m1 = m2, where m1 is the mean in population 1
                    and m2 is the mean in population 2
Assumptions:

         alpha =   0.0500  (two-sided)
         power =   0.9000
            m1 =        0
            m2 =        2
           sd1 =        1
           sd2 =        1
         n2/n1 =     1.00

Estimated required sample sizes:

            n1 =        6
            n2 =        6

. sampsi 0 20, sd(10)

Estimated sample size for two-sample comparison of means

Test Ho: m1 = m2, where m1 is the mean in population 1
                    and m2 is the mean in population 2
Assumptions:

         alpha =   0.0500  (two-sided)
         power =   0.9000
            m1 =        0
            m2 =       20
           sd1 =       10
           sd2 =       10
         n2/n1 =     1.00

Estimated required sample sizes:

            n1 =        6
            n2 =        6

Two standard deviations is a lot. With a two standard deviation  
difference between your means,


. mwstati 2 1 1 , i
Mann-Whitney statistic for this situation is: 0.9214


The Mann-Whitney statistic is 0.92. In other words, there is a 92%  
chance that a person with MS taken at random will have a higher  
cytokyne measurement than a control taken at random. You might  
consider a Mann Whitney statistic of 0.75 to be more realistic as  
indicating a clinically important difference. This occurs with a  
difference of 1 standard deviation.


Ron�n Conroy
rconroy@rcsi.ie




*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/