Face to face with an expert: Exploring joint visual attention during forensic face and feature comparison learning in expert-novice pairs

Visual expertise occurs in professional domains where comprehension of visual material is central to performance, such as in forensic analysis. Despite the prominence of apprenticeship in various visual domains, the didactic strategies employed by visual experts to scaffold novices during one-on-one interactions remain insufficiently explored. This multiple case study introduces a data-driven approach to capturing and coding expert-novice pairs’ interaction based on joint visual attention (JVA). Three registered forensic experts (>10 years’ experience) taught novice psychology students face- and feature comparison tasks (comparison of (passport) photos, and matching security footage to a 3D model). Six sessions were captured with dual wearable eye tracking, alongside video and audio recordings. Results showed that during moments of high JVA experts tend to employ cueing strategies more frequently than during moments of low JVA, although chunking techniques are used equally during moments of high and low JVA. Novices interacted with experts often through cues, but rarely demonstrated chunking, showing a focus on the logistics or “how”, whereas experts approached explanations from the “why”. Novices rarely initiated eye contact, whereas experts often did, for example to check novices’ understanding. The differences in the perceptual and procedural nature of the two comparison tasks led to variations in scaffolding and gaze behaviours. Isolating segments of high and low JVA captured with wearable eye tracking allows for a portable and naturalistic data-driven approach to expert-novice interaction that captures nonverbal actions like gesturing in-camera. This research provides a foundation for various avenues of future research.