VISUAL BEHAVIOUR IN SOCIAL INTERACTION: THE VALIDITY OF EYE‐CONTACT ASSESSMENTS UNDER DIFFERENT CONDITIONS OF OBSERVATION

Abstract

On each session of the experiment, each of five subjects was assigned to act as either gaze-sender (GS), gaze-receiver (GR), direct-sight observer (DSO), TV observer (TVO) or one-way mirror observer (OWO). The GS had to look into the GR's eyes each time he received prerecorded signal through earphones. The GR and the three observers recorded GS's eye-gaze behaviour. Two conditions of distance between GR and GS were included. Results showed that (1) the GR, DSO and TVO presented an equivalent accuracy in their EC recordings as compared with the programmed signals, while the OWO's accuracy appeared significantly lesser; (2) as a concealed observer, OWO was also less efficient than TVO since the agreement of the former with GR's records was significantly lower than that of the DSO, which was not the case for the latter; (3) with increment of distance, OWOs recorded longer EC periods, a tendency which was not found significant among the other observers.