Once ive found my final minimum adequate model, what other options could i use to ensure that all of the variables dropped during the intercorrelation and model selection process are not entirely disregarded which one would be best?