EViews.com

So I have a question about determining significance of the interaction between variables:

A simplified version of my equation is:

Vote percentage = C + β1econ_cost + β2party_affiliation + β3party_affiliation*econ_cost

where econ_cost is a continuous variable and party_affiliation is a dummy variable

The interaction term is not statistically significant (t-value < 2, p-value >0.05), so I'm not sure if adding the interaction term into the model is useful. Someone told me I should do an F-test to determine this, but I'm not quite sure how.

Any help would be much appreciated.

If you're interested in testing b3=0, then the t-test is all you need.

Basically I want to know if including β3 in the equation is worthwhile. β3 is not significant according to either the t-test or the p-test; does this mean that we disregard the interactive variable completely? Or does the fact that including β3 in the new regression increases R^2/adjusted R^2 worthwhile?