Message-ID: <Pine.GSO.4.31.0212101050410.23694-100000@markov.stats>
Date: 2002-12-10T11:52:02Z
From: Brian Ripley
Subject: Variance of a single number
In-Reply-To: <024D6AEFCB92CB47BA1085751D184BB80105F1CD@MBXSRV03.stf.nus.edu.sg>
On Tue, 10 Dec 2002, Adaikalavan Ramasamy wrote:
> Just out of curiosity, can some please explain the following return NA.
>
> x <- 6
> var(x)
>
> y <- c( NA, NA, 10000 )
> var(y, na.rm=T)
>
>
> Unless I am seriously misguided, I believe that the variance of a single
> number (i.e. a constant) should be zero. Thanks.
You are. It's an estimate of the population variance, about which there
is no information at all in one number. Or just look at the formula,
which gives 0/0.