It is the link between two variables in the sense of interdependence between the variables or in the sense of dependence of a variable to the other. There is correlation between the variables, when the variation of one the other also varies not randomly, if it doesn't happen the variables are called uncorrelated. If the link between two variables is of the linear type, that is expressible by a straight line, we can speak indifferently of correlation or collinearity. Collinearity can be measured using various indexes, among which the best known is the correlation coefficient of Bravais-Pearson: which is equal to the ratio of the covariance between the variables divided by the product of the standard deviations of the variables. If the index takes values positive, negative, zero, the variables are said, respectively, directly related, indirectly related, uncorrelated.




Guseo R., Statistica, CEDAM, 2006.

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


Editor: Giuliano DI TOMMASO