Disadvantageously, dropped variables cannot be re-entered into the model, even though they might be significant in later iterations [ 87 ], but in comparison to forward selection or stepwise models, backward selection does not require strict preselection and can better identify useful combinations of predictors without requiring individual explanatory value of each variable.
← all excerpts
Psychophysical predictors of experimental muscle pain intensity following fatiguing calf exercise.
1
—
—