Unveiling what is written in the stars: Analysing explicit, implicit, and discourse patterns of sentiment in social media

Although researchers often use automated sentiment analysis in assessing online consumer product evaluations by counting positive and negative words, more granular sentiment expressions—such as activation levels, implicit meanings, and patterns of sentiment across sentences (in an online review for example)—are relatively poorly understood. These granularities aid in differentiating different degrees of sentiment strength and enable a more in-depth analysis of the relationship between sentiment expression in consumer verbatim comments and subsequent online behaviour.

Using Speech Act Theory as an enabling framework, this study conceptualised the differential impacts of explicit sentiment force, implicit expressions, and discourse patterns on overall sentiment strength, for example star ratings.

Project outputs

We demonstrated the significance of these conceptualisations in an empirical study using online consumer reviews, as well as two follow-up studies assessing their relevance for sales and generalisability across social media contexts. By focussing on how consumers express different degrees of sentiment strength, this study offered a more in-depth understanding of online consumer behaviour.