I'm testing 2 possible independent continuous variables as predictors of the same continuous dependent variable. Both of them are measures of obesity.
The Loess regression for one of them shows a pretty linear tendency, while the other one is closer to a sigmoid shape.
Can I state that the one with a linear tendency is as best choice as predictor of the dependent variable in a clinical setting if both of them have a similar goodness of fit? Where can I find a reliable reference for that?
Bart Kemper
This depends on the sensitivity and criticality of the ends. For example, if there is a failure event or other critical aspect in the ends, a linear assumption will not be appropriate.
0 votes
0 thanks
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression
More Caio Alves's questions See All
Similar questions and discussions