A statistic that provides an indication of the extent to which a multiple regression model can be generalized. This is achieved by fitting the model to every subset of n−1 of the n observations, y₁, y₂,…, y_n, of the response variable (see regression). Let be the fitted value for y_j based on the model that uses all the observed values except y_j. The corresponding residual ε̂_−j is given byThe PRESS statistic iswhere is the mean of the n observations. The abbreviation PRESS is derived from predicted residual sum of squares. The statistic is an analogue of the R² statistic (see anova) and values close to 1 are preferable to those close to 0.