stepwise entry

a method in which predictors are entered according to some statistical criteria in statistical methods such as logistic regression, multiple regression, discriminant function analysis and log-linear analysis. The results of such an analysis may depend on very small differences in predictors meeting those criteria. One predictor may be entered into an analysis rather than another predictor when the differences between the two predictors are very small. Consequently, the results of this method always need to be interpreted with care.