There are generally two types of standard deviation calculations, one for when your data represents a sample of the population, and one where it represents the whole population. The former divides the squared deviations by N-1 (the sample count minus 1), whereas the latter divides by N.
As pryrt points out, you can get Statistics::Basic to do both by setting UNBIAS to true, whereas Statistics::Lite gives you two separate functions. Note also that Statistics::Basic defaults to a precision of 2 decimal points - this can be changed by setting IPRES to the desired precision.
Just for fun I did a little comparison table:
| Excel |
STDEV |
0.41068873 |
STDEVP |
0.400791162 |
| Open Office |
STDEV |
0.4106887305 |
STDEVP |
0.400791162 |
| Statistics::Lite |
stddev() |
0.410688730467897 |
stddevp() |
0.400791162022762 |
| Statistics::Basic |
stddev()
UNBIAS true |
0.410688730467897 |
stddev() |
0.400791162022762 |
