I believe davies was wrong about it also being called the standard error: according to wp, Standard_error, the standard error is the standard deviation divided by the root of the number of samples.
The population standard deviation (Excel: =STDEVP() or =STDEV.P()) assumes you that have data for the entire population; the sample standard deviation (Excel: =STDEV() or =STDEV.S()) assumes that your n data points are just a sample of the population, so it uses a different denominator (n-1) inside the radical, to account for the uncertainty because it's just a sample, not the whole population. Given your set of data, the 0.401791 is the population standard deviation, and the 0.410689 is the sample standard deviation.
According to Statistics::Basic, you can set $Statistics::Basic::UNBIAS true to get the sample, or (implied) set it false to get the population. Statistics::Lite directly says that their stddev() function is the unbiased (sample) n-1, which matches the Excel =STDEV() function, and they have the stddevp() for the population n, matching the Excel =STDEVP().
In reply to Re^3: Excel or Perl for simple statistics?
by pryrt
in thread Excel or Perl for simple statistics?
by Anonymous Monk
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |