Actually, I dont' think that's correct

[Gene]

The formula for standard deviation of a sample has n-1 on the denominator, so it should work for 2 data points.

Suppose the data are 10 and 5

Standard deviation is

SQRT ((( 10 - 7.5 ) ^2 + (5 - 7.5) ^2) divided by (2 - 1))

Which works out to be 2.5

Granted, that's not exactly helpful with only 2 data points, but it does work.

Population standard deviation only has n on the denominator, so you can compute it when you only have 1 data point...the value of the standard deviation in that instance would be 0.