Eyes wide apart : Overestimating interpupillary distance À

Our ability to perceive many aspects of the human face is so acute that some would propose that we possess specialised face-processing mechanisms. One example of this exquisite performance is our ability to discriminate differences in the lateral separation of the eyes in pairs of otherwise identical facesösee Kemp et al (1990, Perception 19 531 ^ 543). However, we have found that our ability to estimate our own interpupillary distance (IPD) shows no such accuracy. Subjects were asked to indicate, without undue rumination, their own IPD and, by way of control, the distance from the bridge to the tip of their own nose. Compared to the true distances, IPD was overestimated by more than 30% while nose length was underestimated by 11%. Although a plausible reason for the nose length underestimation is foreshortening of the image, the reason for the overestimation of eye separation remains obscure. DOI:10.1068/p3350 À This title has been chosen in homage to Michael Morgan who wrote a piece in The Guardian newspaper entitled ``Eyes wide apart'', in which some aspects of the findings of this study are discussed. (1) It should be stressed here that the terms `positive' and `upright' refer to contrast polarity and orientation and not to any moral quality. The face used in this experiment was that of Michael Morgan. 2 Methods 2.1 Subjects Eighty-one undergraduates (eighteen males, sixty-three females) took part in the experiment. All might be described as `Caucasian' but a more accurate description would be `white European'. All subjects were na|« ve to the purpose of the experiment. 2.2 Design All subjects estimated their own IPD and the length of their own nose. The actual IPD and nose length of each subject was measured by the experimenter. 2.3 Materials and apparatus Subjects were presented with a response sheet (see figure 1) and received detailed instructions as to their task. First, they were informed that they should estimate their IPD, ie the distance from the centre of the pupil of one eye to the centre of the pupil of the other eye. The position of the centre of one eye was already marked on the upper part of the response sheet with a vertical tick-mark; their task was to draw a second tick-mark that delineated the IPD. Second, they were asked to rotate the response sheet through 908 and mark the distance from the bridge to the tip of their nose on the second (now vertical) line. Instructions were given that the bridge of the nose was defined as the point on the midline of the nose intercepted by a line between the pupillary centres. The definition of the tip of the nose was less rigorous, being simply `the tip of the nose'. All subjects reported that they understood the instructions unambiguously. One week after these estimates had been made, the real IPD and nose length of each subject was recorded. Two different IPD measurements are commonly described in the literature: first the `near IPD' measures the IPD when the subject is fixating a point 14 inches away, while the `distance IPD' measures the IPD with the subject's visual axes parallel, that is fixating a distant object. The near IPD is generally 2 ^ 3 mm less than the distance IPD; it is not obvious whether, asked to estimate their IPD, subjects would choose one of these measures rather than the other but, given that our hypothesis in this paper is that perceived IPD is overestimated, the distance IPD value was measured. The measurement was carried out along the lines of the standard manner recommended in optometry and ophthalmology texts (eg Reinecke and Herm 1983, page 173). The subject, facing the experimenter, was instructed to look at the experimenter's left-eye pupil; the experimenter looked into the subject's right eye with his left, and aligned the zero value of a millimetre rule with the centre of the subject's right pupil. The subject Figure 1. The response sheet. Subjects were asked to mark their IPD on the upper line; for half the subjects the already-marked point was on the left, for the remaining subjects this point was on the right. Having estimated IPD, subjects were invited to rotate the response sheet through 908 and indicate on the lower, now vertical, line their nose length. 652 P Thompson