A computationally intensive measure of the strength of the relationship between two random variables. The measure, proposed in 2011, makes use of the mutual information (see bivariate distribution) between the variables. A sequence of lattices is superimposed on a scatter diagram showing the observed pairs of values. For each lattice, the mutual information value is scaled by division by the maximum possible mutual information value for that size of lattice. The MIC is then defined as the maximum over these scaled values. A strength of the procedure is that it can be used with any type of relationship (which might consist of a mixture of several functions), whereas the correlation coefficient refers only to linear relationships.