Assuming I have a group of 'A' data with mean =0.852361, sd=0.388232 and sample size=34. And I have another group of 'B' data with mean = 0.798698, sd=0.396084 and sample size=40. Now I want to calculate the compound error of (A-B)/A*100. To calculate the total compound error, I first calculated the compound error of (A-B) by SQRT((SD of A)^2 + (SD of B)^2) and then calculate the total compound error of (A-B)/A*100 by (A-B)/A*SQRT(((SD of A)/A)^2 + ((SD of (A-B))/(A-B))^2)*100. So the total calculated compound error is SD (standard deviation)= 65.13214. Is it right?

Then I want to know what is the sample size for the value of (A-B)/A*100? If A and B have the same sample size, I think the value of (A-B)/A*100 should have the same sample size. But here A and B have different sample size, I don't know how to estimate the sample size of the final value. Currently, I calculated the SEM (standard error of the mean) of A and B by (SD of A or B)/SQRT(sample size) and then calculated the compound error of SEM, and got the SEM of (A-B)/A*100 as 10.73517. And then calculated the sample size of (A-B)/A*100 by (SD/SEM)^2, which is 36.81061. Is it correct?

Thanks very much if anyone can help......

Best,

WZ