ABSTRACT
Statistical sophistication, at any level, is at the mercy of the data actually provided. Measured predictors that are in fact dependent on the outcome will not be useful in predicting the outcome. Well-organized data not only expedites analyses but also eliminates situations where the analyst may have to (mis)interpret the data. Do not attempt to force predictors or outcomes into categories for convenience, although it may be necessary for presenting tables or graphs.
