Suppose the sample has [tex]n[/tex] data points, each denoted [tex]x_i[/tex] where [tex]1\le i\le n[/tex].
First compute the mean: [tex]\overline x=\displaystyle\frac1n\sum_{i=1}^nx_i[/tex].
The sample standard deviation is then
[tex]\displaystyle\sqrt{\sum_{i=1}^n\frac{(\overline x-x_i)^2}{n-1}}[/tex]
