The effect of an experimenter's presence on the quality of interpretable data in human figure drawings