Covariance Calculation

  • Thread starter Thread starter Guest
  • Start date Start date
G

Guest

When excel calculates covariance, why does it use N as the denominator instead of the standard N-1? Is this my ignorance of statistics working here, is there a valid reason, or is it a bug in the formula

Thanks

LK
 
When excel calculates covariance, why does it use N as the denominator
instead of the standard N-1? Is this my ignorance of statistics
working here, is there a valid reason, or is it a bug in the formula?
Click to expand...

Most likely it is trying for compatibility with some long forgotten
competitor.

You can of course multiply by N/N-1 if you need the standard.
 
Most likely it is trying for compatibility with some long forgotten
competitor.
Click to expand...

Got any specifics to go with that thoroughly ignorant shot-in-the-dark?

From 123's online help:

"@COV

@COV(range1;range2;[type]) calculates either the population or sample covariance
of the values in range1 and
range2.

Arguments

range1 and range2 are the names or addresses of ranges. range1 and range2 must
be the same size. If range1 or
range 2 are not the same size, @COV returns ERR.
1-2-3 pairs cells in the two ranges by their same order in the range. Ranges are
ordered from top to bottom, left to
right, first sheet through last.
type is an optional argument that specifies whether to calculate the population
or sample covariance.

type 1-2-3 calculates

0 Population covariance; default if you omit the argument
1 Sample covariance"


Perhaps you're referring to some shareware spreadsheet you heard someone telling
someone else had a covariance function that worked like Excels.

To address your evident ignorance, Excel had its COVAR function before 123 had
its @COV function, so Microsoft screwed this up all on their own.
 
Back
Top