25
$\begingroup$

This is my first question on Cross Validated here, so please help me out even if it seems trivial :-) First of all, the question might be an outcome of language differences or perhaps me having real deficiencies in statistics. Nevertheless, here it is:

In population statistics, are variation and variance the same terms? If not, what is the difference between the two?

I know that variance is the square of standard deviation. I also know that it is a measure of how sparse the data is, and I know how to compute it.

However, I've been following a Coursera.org course called "Model Thinking", and the lecturer clearly described variance but was constantly calling it variation. That got me confused a bit.

To be fair, he always talked about computing variation of some particular instance in a population.

Could someone make it clear to me if those are interchangeable, or perhaps I'm missing something?

$\endgroup$
  • 5
    $\begingroup$ Variation, unlike variance, is not the name of some specific quantity (however, Coefficient of variation is). It is a generic term, like variability. It is just amount of variability which can be measured by various quantities (most popular of them being variance). $\endgroup$ – ttnphns Mar 1 '14 at 7:36
  • $\begingroup$ So basically you are saying that Variance is a real statistical term with a formal model standing behind it, but variation is just a word describing relation between expected & real data? $\endgroup$ – ŁukaszBachman Mar 1 '14 at 7:42
  • $\begingroup$ Right - I changed that :) $\endgroup$ – ŁukaszBachman Mar 1 '14 at 7:46
  • 2
    $\begingroup$ Variance has a formula. Variation has no one formula, it is a generic term. Both variance and variation can be 1) a statistic describing a sample, 2) a parameter describing a population, 3) a statistic as an estimate of the correstonding parameter $\endgroup$ – ttnphns Mar 1 '14 at 7:46
  • 4
    $\begingroup$ Another analogue here is "spread." There isn't a formal equation for calculating "spread," although it's appropriate to say that "variance" is a measure of "spread." I think in this context "spread" and "variation" are equivalent. $\endgroup$ – David Marx Mar 1 '14 at 8:25
19
$\begingroup$

Here's a full wikipedia article discussing this topic: http://en.wikipedia.org/wiki/Statistical_dispersion

As described by others in the comments here, the short answer is: no, variation $\ne$ variance. Synonyms for "variation" are spread, dispersion, scatter and variability. It's just a way of talking about the behavior of the data in a general sense as either having a lot of density over a narrow interval (generally near the mean, but not necessarily if the distribution is skewed) or spread out over a wide range. Variance is a particular measure of variability, but others exist (and several are enumerated in the linked article).

$\endgroup$
11
$\begingroup$

@ttnphns is right, but since the info wasn't written as an answer, I'm going to attempt to steal the credit! :)

Variation may be understood best as a general term for a class of different concepts, of which $(\sigma^2)$ is only one. Levine and Roos (1997) also consider $(\sigma)$ a variation concept, among others.

To demonstrate why the distinction might be important, compare also the $(\frac\sigma\mu)$, and the mathematical concept, total variation, which has several definitions unto itself. Then there are all manners of qualitative variation, which are mentioned in the Wikipedia article @DavidMarx linked. These pages corroborate his answer BTW; statistical dispersion or variability are better synonyms for variation than variance, which is clearly not so synonymous.

BTW, here's a cool GIF of one kind of total variation: the length of the path on the $y$ axis that the red ball travels. Definitely not the same as variance!

Reference
Levine, J. H., & Roos, T. B. (1997). Description: Numbers for the variation. Introduction to data analysis: The rules of evidence (Volume I:074). Dartmouth College. Retrieved from http://www.dartmouth.edu/~mss/data%20analysis/Volume%20I%20pdf%20/074%20Description%20%20Numb%20for.pdf

$\endgroup$

protected by Glen_b Nov 25 '16 at 1:00

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.