For me thesis I have a sample of compiling times in Android, however, some people says the standard deviation is calculated by this method: $$\sqrt{\frac{\sum (x-\bar x)^2}{(n-1)}}.$$ Other methods divide by the # of tests. What is the difference between the two? Sorry for my bad English.
$\begingroup$
$\endgroup$
1
-
$\begingroup$ Formatting tips here. $\endgroup$– Em.May 17, 2017 at 20:55
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
The difference is whether you would like to use a biased (with $n$) or non-biased (with $n-1$) estimator. Please see Wiki on unbiased estimation of standard deviation for detailed info.
