I have seen a previous interesting post about how to test the significance of a categorical variable in linear regression How to test the statistical significance for categorical variable in linear regression?
My question is when it is recommended in a linear regression to do a statistical significance test of the categorical variable (instead of one p-value for each category/dummy variable)? Is there specific situation in which you need to have an overall significance test (or it is really your preference )? If yes, what if you are reporting confidence intervals, do you still need to report an overall CI for the categorical variable? Thanks
