Statistical estimate of the extent to which a score distribution, or other numerical series of observations, differs significantly from the numerical values predicted by a mathematical model (i.e., an estimation of how well the theory fits the data)