One predictor overshadows all other?

[marco_c88]

Hi everyone! I was wondering, how do I deal with a predictor that overshadows all others? All predictors included in the model are statistically significant, but one of the has a very strong impact on the dependent variables, so much that when its value is high, the impact of all other predictors is virtually non-existent. How do I deal with this? Thank you very much!

 

[hlsmith]

Do you know the underlying context pinning your modeling question? Perhaps it is just the best term and the others are poorly associated with the outcome. What would be the problem with that occurring. In addition, it may be just a more informative version of the other variables.

 

[marco_c88]

Do you know the underlying context pinning your modeling question? Perhaps it is just the best term and the others are poorly associated with the outcome. What would be the problem with that occurring. In addition, it may be just a more informative version of the other variables.

It is definitely the best predictor, true. However, the other predictors do have an impact, but only when that one predictor is at low values...

 

[hlsmith]

Define low values and overall context. Examine variance inflation factors and concept of partial residuals.

 

[marco_c88]

Define low values and overall context. Examine variance inflation factors and concept of partial residuals.

Thanks!