Face processing
What do face masks do to how we read faces?
Open access · cc by · source: Europe PMC
Surgical masks made adults worse at naming celebrities, reading most emotions and judging gender, but they actually made angry faces easier to recognise.
Study at a glance
- Design
- Human experiment — Online within-subject experiment: every participant did four face tasks with intermixed masked and unmasked faces, with participant gender as a between-subjects factor
- N
- N=108 · 108 adults analysed (56 men, 52 women) after excluding 2 of 110 who finished implausibly fast; the age-estimation analysis appears to have slightly fewer cases (its degrees of freedom drop by one)
- Population
- Adult online participants recruited through Testable Minds; the authors describe the sample and face stimuli as Caucasian
- Outcome
- Famous-face naming accuracy, six-choice emotion categorisation accuracy, gender classification accuracy, and absolute age-estimation error, for masked versus full faces
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
Masks lowered famous-face naming accuracy (from 0.65 to 0.61 on average), mainly for female celebrities. Emotion recognition dropped sharply, from 61.77% correct on full faces to 44.97% with masks, for happy, sad, surprised, disgusted and fearful faces; anger was the exception and was recognised better when masked. Gender classification fell modestly, almost entirely for female faces (about 92% to 85%), while masks had no overall effect on age error, though female faces' ages were estimated less accurately when masked. Women outperformed men at emotion recognition, and men's famous-face naming was hurt by masks while women's was not.
Methodology
Adults recruited online completed four tasks in which half the faces had a surgical mask digitally added over the nose and mouth. They named 20 famous faces, picked which of six emotions a face showed, classified hairless faces as male or female, and typed in the estimated age of faces from their 20s to 70s. Masked and unmasked trials were mixed within each task, and the authors also compared male and female participants.
Limitations
Faces were photographs with masks pasted on digitally and viewed briefly online, which differs from meeting a masked person with voice, movement and context. Masked and unmasked faces were intermixed, and the authors note this may have encouraged picture-matching strategies that make mask effects look smaller than in a blocked design. The authors say participants and faces were all Caucasian, so cultural generalisation is untested. Several of the participant-gender interactions were only just significant and their direction was not firmly predicted, so they need replication.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
Not yet placed on a claim. This paper has study layers, but no concept page yet cites it as support, challenge, or qualifier.
Related papers in this topic
Same topic cluster — not a recommendation engine.