It is a measure of joint variability of two variables X and Y.

Where X, Y are the observed values of the two variables and E[X], E[Y] are their arithmetic means.

The covariance is the average of the products and therefore is positive in the case of concordance, that is, if a positive or negative difference of X by E[X] correspond mainly difference of equal sign of Y by E[Y]. Conversely, it is negative in the case of a discrepancy between the signs of the differences between the two variables. The link between the two variables is defined direct in the first case, and reverse in the second case.


Guseo R., Statistica, CEDAM, 2006.

Leti G., Statistica descrittiva, Il Mulino, 1999.

Editor: Giuliano DI TOMMASO