Using Eye-Tracking and Retrospective Think Aloud as a Probing Tool in Food Labelling Research: An Abstract

Abstract

Subconscious behaviours can have a profound impact on decision-making and purchasing outcomes, yet can go unreported using self-reported measures. In the labelling domain, self-reported measures are subject to over-reporting and misreporting (Grunert et al. 2010; Graham et al. 2012), thereby leading to increased interest in eye-tracking (ET) applications in this domain. However, this has resulted in an over emphasis on label-specific factors, leading to calls to situate label usage within the broader decision-making context (Mawad et al. 2015). Given the efficacy of self-reported measures over ‘objective’ measures in capturing the motivational component of label usage (Miller et al. 2015), this paper argues that, although paradigmatically opposed, supplementing traditional self-reported introspective techniques with ‘objective’ (ET) behavioural measures can enhance probing, offering new insights into consumer behaviour. This study adopted a sequential data collection approach leveraging ET methodology, retrospective think aloud (RTA) protocols and qualitative interviewing. Yogurt acted as the case food for this study. Seventeen participants were recruited using quota-based purposive maximum variation sampling. Participants were assigned an open-ended task to avoid biasing of gaze data through atypical information search, and were presented with labels for five fictitious yogurt products, based on an audit of contemporaneous labels. A remote eye-tracking system was located beneath a computer monitor, and participants’ progress during the experiment was monitored remotely for real-time visual (qualitative) analysis of label usage data, which facilitated preparation of resources for the post-experiment interviews. No time limit was placed on the viewing. ET outputs were explored through a cued RTA protocol and qualitative interviewing. RTA allowed participants to retrace the sequence of cognitive events enacted during the ET experiment without creating the increased cognitive burden experienced with concurrent think-aloud (Elling et al. 2011). Discrepancies and omission between ET and RTA data were recorded and acted as the basis for probing during the subsequent interview, allowing for more precise questioning, reflective of individual participants’ interaction with the experimental labelling stimuli. Various label attributes reported as salient in decision-making during interviewing were not viewed during ET experiments, indicating discrepancies between reported and observed behaviours. Interviewing revealed gaps between reportedly important attributes and motivational drive to seek out attribute relevant information. Consequently, qualitative analysis of ET visualisations prior to introspective data collection was demonstrated to improve probing. Future research may consider the role of introspective techniques in supplementing ET research in better accounting for the factors underlying ET observations.