The effects of sampling artifacts on cosmic dust flux estimates: a reevaluation of nonvolatile tracers (Os, Ir)