Statistics Question: should I throw out and independent variable it it has little impact?

I am preforming a regression analysis where four parts of a test are used to determine the impact on an employees future success. In stating my results, should I re-run the analysis and leave out the independent variable (which has very little impact on the success rating) and provide those findings also?Statistics Question: should I throw out and independent variable it it has little impact?
Are you using any model selection criterion? What the AIC or AICC values for the different models with and without the variable? Is the variable significant in an interaction? All these questions need to be addressed before you remove the variable from the model.



You can also conduct an F test for the reduced model vs. full model, here is a link to that information:



http://www.public.iastate.edu/~pcaragea/鈥?/a>





here is link to model selection criterion http://en.wikipedia.org/wiki/Akaike_info鈥?/a>