Do people from different cultures look at faces differently?
Western and East Asian observers recognised faces equally well but gathered the information differently: Westerners fixated the eyes and mouth, while East Asians fixated the centre of the face and used more peripheral vision.
Source
Mapping Face Recognition Information Use across Cultures
Study at a glance
- Design
- Cross-sectional — Lab comparison of two cultural groups performing an old/new face recognition task under a novel gaze-contingent 'Expanding Spotlight' that reveals the face only around fixation, growing with fixation duration.
- N
- N=30 · 15 Western Caucasian students (Glasgow) and 15 East Asian students (Guangzhou); analyses used correctly recognised trials.
- Population
- Young adult university students in the UK and China with normal or corrected vision.
- Outcome
- Recognition accuracy and reaction time, global eye-movement measures, and statistical fixation maps of where each group looked on the face.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
Fifteen Western Caucasian and fifteen East Asian students learned faces of both ethnicities and then judged whether faces were old or new, with facial expressions changed between learning and test to prevent simple picture matching. Eye movements were recorded with a gaze-contingent 'Expanding Spotlight': the real face was visible only through a window around the current fixation that grew the longer the person kept looking, with an average face shown elsewhere. The authors built fixation maps for each group and reconstructed the visual information each group actually took in.
What they found
Accuracy, reaction times and overall eye-movement measures (number and length of fixations, saccade length) did not differ between groups. Western observers spent longer fixating the eyes and mouth (286 vs 106 ms), whereas East Asian observers spent longer on the centre of the face (368 vs 100), with large effect sizes (Cohen's d of 0.92 and 1.06). Reconstructed information maps showed the groups took in different visual input, which the authors argue is hard to reconcile with a single, universal holistic process for faces.
The limits
What it doesn't show
The groups were small (15 per culture) and were tested at different universities, so differences could partly reflect site, language or other group factors rather than culture alone. The gaze-contingent window changes normal viewing, and the authors concede that tying fixation duration to how much face is revealed could inflate peripheral information. Eye movements show where information was sampled, not what mental representation was formed, so the study cannot prove whether processing was holistic or featural. Culture was not manipulated, so the causes of the bias (e.g. experience or social norms) are not tested.
Key terms
- Gaze-contingent display
- A screen that changes in real time depending on where the viewer is looking, used to control what information is visible at each fixation.
- Holistic processing
- The idea that faces are perceived as integrated wholes rather than as separate features; the authors argue the term is poorly defined.
- Foveal vs extrafoveal vision
- Foveal vision is the sharp central few degrees of the visual field; extrafoveal vision is the lower-resolution periphery.
- Fixation map
- A spatial map showing where and for how long observers fixated on an image, often compared statistically between groups.
- Old/new recognition task
- A memory test in which participants study items and then judge whether each test item was studied (old) or not (new).
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
Where did East Asian observers mainly fixate during face recognition?
Common questions
If East Asian observers look at the nose, do they ignore the eyes?
No. Previous work by the same group shows both cultures use the eyes and mouth; East Asian observers appear to sample them from a central fixation using peripheral vision.
Why change the facial expression between learning and test?
So participants could not succeed by matching identical images and had to recognise the person's identity.
Does a central fixation pattern prove holistic processing?
The authors argue it does not; it only shows that information was probably picked up away from the fixation point, which could still support feature-based processing.
More on Face processing