I'm doing a multiple regression analysis and comparing it to the simpler model. But I've no idea what the simpler model is.
In the full model, I have a variable that is statistically insignificant in the full model. And in the same model there is an interaction between this same variable and a covariant, and they are highly statistically significant.
But when I'm doing the simple model, do I remove the variable that has no signficance, or do I remove the interaction? It may be a dumb question, just that since it took me ages to begin to understand interactions, I always think of them as "complex" so I figure "well I should just remove it for the simple model" but logically, I think the simple model should remove the factors that aren't signficant.