If you've fired 3 five shot strings, you have a sample size of 15, good 'nuf. Just do your SD calculation on all shots fired. One problem, if they're shot on different days at different atmospheric conditions, it could affect the velocities.
I was mearly thinking of the statistical variance of a 5 shot string vs 10 or more shots. Just that the larger the sample size, the more consistent the variance and therefore the the more accurate standard deviation.