covariances versus covariation

February 10, 2011 10:44 AM

Okay, stats mavens: What's the difference between covariance and covariation?

Wise up a chump. Give a sucker an even break. What is the difference?

posted by internet_explorer to Computers & Internet (5 answers total)

what makes you think there is a difference? I've never heard them used as two different things, and the top hit for googling "covariation" is the wikipedia page for "covariance".
posted by milestogo at 11:10 AM on February 10, 2011

To my knowledge, covariance is the statistical term, whereas covariation is just a synonym to describe to variables that vary together or the amount to which their variation is shared.
posted by proj at 11:33 AM on February 10, 2011

I echo proj. A covariance is a specific measurement -- the expected value of [(x-xbar) times (y-ybar)].* Covariation is the general term for things going together (or against each other).

*If you're curious or that was your next question, a correlation is a covariance that's been rescaled in a particular way and lacks dimensions or units.
posted by ROU_Xenophobe at 11:52 AM on February 10, 2011 [1 favorite]

I agree with the above. I would say that covariance is the statistic that describes covariation.
posted by pemberkins at 11:58 AM on February 10, 2011

To emphasize the above, the wiki has a nice plot of variables that are not independent, but do have zero covariance.
posted by a robot made out of meat at 6:27 PM on February 10, 2011