Methodological issues about interobservation: An example from an attraction effect study