How Do Pictures Shape "Liking" Behavior? A Stimulus-Organism-Response Perspective

In recent years, online content has shifted from text-centric to visual-oriented content. Despite the massive amount of pictures embedded in online content, how pictures affect people’s liking remains understudied. Using the Stimulus-Organism-Response (SOR) model, the current work investigates this question in the context of the catering industry. We divide the model into two linkages: a stimulus-organism linkage, in which pictures acting as stimuli influence viewers’ affective and cognitive states, and an organism-response linkage, in which such states shape “liking” behavior. Picture features are further classified into domain-related and visual aesthetics features. The affective state is measured by valence and arousal according to the dimensional model of emotion. The cognitive state is measured by perceived calorie content. Using data from dianping.com, one of the largest online review platforms in China, we combine manual tagging and machine learning approaches to measure picture-evoked affective and cognitive states. Leveraging a hierarchical generalized linear model, we find that the domain-related and visual aesthetics features of pictures significantly affect valence, arousal, and perceived calorie content, further influencing liking behavior. This study demonstrates the effect of emotions evoked by visual information, thus providing managerial implications for review providers, platform managers, and restaurant owners