Opening the black box. Three logics of data interpretation in participant observation studies and their consequences in research