search

commfish / southeast_pink_salmon_preseason

Preseason pink salmon forecasts.
0 stars 1 forks source link

Interactions #3

Open fssem1 opened 4 years ago

fssem1 commented 4 years ago

https://newonlinecourses.science.psu.edu/stat462/node/157/

Typically, regression models that include interactions between quantitative predictors adhere to the hierarchy principle, which says that if your model includes an interaction term, X1X2 and X1X2 is shown to be a statistically significant predictor of Y, then your model should also include the "main effects X1 and X2, whether or not the coefficients for these main effects are significant. Depending on the subject area, there may be circumstances where a main effect could be excluded, but this tends to be the exception.

rich-brenner commented 4 years ago

Just so I'm clear on this principle. A valid model would be: Y = X1 + X2 + X1*X2

Whereas the following would not be valid based on the hierarchy principle: Y = X1*X2 ??

fssem1 commented 4 years ago

I think that could still be a valid model if you had justification (i.e., an exception). My main concern was that one of the models brought forth was: Y=X1 + X2*X1 I think there also needs to be an X2 coefficient in the above model.

rich-brenner commented 4 years ago

And more: https://www.quora.com/Why-do-we-have-the-hierarchical-principle-in-adding-interactions-to-a-model-What-is-the-significance-of-it