Gather in data inputs for the power test.
Determine the acceptable rate of error, which is the "t" statistic or alpha. A 5 percent error rate would equal an alpha of .05.
Find the mean and standard deviation of your population set. The mean is simply the average whereas the standard deviation is the number of points a typical measurement is away from the mean. Assume the mean is 20 and standard deviation is 3. Count this size of your population; assume that it is 50 samples.
Enter the following statements into the SAS program.
proc power;
onesamplemeans
mean = 20
ntotal = 50
stddev = 3
power = .;
run;
Run the program through SAS, which will output a power statistic. The power statistic is a percentage term that identifies the probability that your sample size is large enough. Your power statistic should be above .95 for considerable accuracy of the sample size.