a transformation that enables the relationship between one or more predictor variables and an outcome variable to be described by a summative formula and thus to be plotted by a straight line when graphed. Based on maximum likelihood estimation, the technique transforms the outcome variable to obtain linearity and approximate normality in a data set. Compare Box–Tidwell transformation. [George E. P. Box (1919–2013) and David Roxbee Cox (1924– ), British statisticians]