I am looking to do a random sampling analysis of a case/control dataset containing 6X as many controls as cases. Therefore, I need to correct for this overabundance of controls (without simply removing the controls) using disproportional sampling.
Is switching the # cases and controls a valid approach, for example if I have 1000 cases and 6000 controls, choosing a random 6000 as cases and 1000 as controls?