Relative importance of metrics in linear regression

Question

I am trying to compute the relative importance of regressors using the relaimpo package. In the linear model, some regressors are insignificant (p-value greater than 0.05). Should I include these regressors while computing

Answer 1

I don't know aything about how the relaimpo package works and neither am I sure about what you exactly mean by "relative importance of regressors", but you must definitely include every "non-significant" regressor in your calculations if they are in the model.

You can also remove those regressors completely, but in that case, fit a new model without them, because the importance of the rest of regressors may see some changes, and whatever "relative importance of regressors" means, it should talk about the model you actually have between your hands