With not significant regression intercept, do I increase the independent variables or the interaction effect variables if the sample size if fixed?
I have a logistic regression model output where only the intercept variable is not significant while all other independent variables are significant.
If I have a fixed number of sample, shall I increase the independent variables or add more interaction effect variables to raise the intercept's significance level?
Thank you very much in advance.
If the intercept is not significant, this means that is not much important and there is no need to make it important on purpose. Even increase the independent variables, may not make it significant. But, if increase the independent variables could increase the accuracy of the module, you may try it.
What about including interaction variables? I tried this and it made the intercept significant.
The accuracy of a model is the most important. The intercept’s significance level does not affect the accuracy. So it is not important if the intercept is significant or not.