"Sample size was calculated on the basis of the primary endpoint, assuming a SD of 20%, a 2-sided type I error of 5%, a power of 80%, and a difference of 20 points in percentage change from baseline.."
How to calculate sample size with 20% SD.. Is it 20% of mean, which is not included in the paper?
Here's what I think is going on. The paper is discussing how they determined sample size. The formula for sample size for a two-sample t test involves Cohen's d. Cohen's d is essentially the difference in means divided by the pooled standard deviation of the samples. You can get here from starting with expected means and expected standard deviations. These standard deviations can be expressed as a percentage of the expected means. For example, I took the example from the bottom of the following page. (http://rcompanion.org/rcompanion/d_02.html). (Caveat, I am the author of the page.) You can run this R code at https://rdrr.io/snippets/. In this example, there is a difference in the means of 3, and the sd is ~ 5% of the mean.
M1 = 100.6 # Mean for sample 1
M2 = 103.6 # Mean for sample 2
S1 = 5.26 # Std dev for sample 1
S2 = 5.26 # Std dev for sample 2
Cohen.d = (M1 - M2)/sqrt(((S1^2) + (S2^2))/2)
library(pwr)
pwr.t.test( n = NULL, d = Cohen.d, sig.level = 0.05, power = 0.90, type = "two.sample", alternative = "two.sided")
Salvatore S. Mangiafico That was really helpful.. Thanks a lot!
David Eugene Booth Thank you!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size