For introductory statistics, should n be used as
the divisor for the sample variance?
Terry
Speed explores the age-old question every statistician is asked to answer at
some point in their career….
“Why is
the denominator in the sample mean n, but the denominator
for the sample variance is n−1?”
Do
you start talking about sample vs population, parameters vs statistics, bias,
or a personal favorite.. “because you already spent one degree of freedom estimating
the mean.” Either way, it’s never a short,
routine answer and the explanation might depend on the audience asking the
question. Then again, maybe in this day
and age, just get a big enough sample and it doesn’t really matter.
http://bulletin.imstat.org/2012/12/terences-stuff-n-vs-n-1/
