When the predictor units of measurement differ greatly in magnitude, which might be useful?
A)Stepwise regression to detect which predictors are best.
B)Decimal transformations to improve data conditioning.
C)Logistic regression to produce probability predictions.
D)Best subsets regression to find the best model.

Question

Quizplus · Accepted Answer

Decimal transformations can help improve data conditioning by bringing the units of measurement closer in magnitude, which can make it easier to compare and interpret the data. This can be particularly useful in regression analysis where the magnitude of the predictors can influence the results. Stepwise and best subsets regression may also be useful for variable selection, but they don't address the issue of differing units of measurement. Logistic regression may be useful if the outcome variable is binary, but it doesn't specifically address the issue of differing predictor units.

When the Predictor Units of Measurement Differ Greatly in Magnitude