An Unbiased Variance Estimator with Denominator $N$

Dai Akita

Standard practice obtains an unbiased variance estimator by dividing by $N-1$ rather than $N$. Yet if only half the data are used to compute the mean, dividing by $N$ can still yie…