Skip to content

Variance of a single number

3 messages · Adaikalavan Ramasamy, PIKAL Petr, Brian Ripley

#
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.

Regards, Adai.
#
Most probably:


     The denominator n - 1 is used which gives an unbiased 
estimator of
     the (co)variance for i.i.d. observations. These functions return
     `NA' when there is only one observation (whereas S-plus has 
been
     returning `NaN'), and from R 1.2.3 fail if `x' has length zero.

from help page for var

try
?var
On 10 Dec 2002 at 15:50, Adaikalavan Ramasamy wrote:

            
Petr Pikal
petr.pikal at precheza.cz
p.pik at volny.cz
#
On Tue, 10 Dec 2002, Adaikalavan Ramasamy wrote:

            
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.