You are here

Kimberly Legocki

DBA Graduate - 2019

Thesis title

Exploring eWOM sentiment and volume on reputation scores: a one year study of 75 global brands on Twitter


Kristen Walker
Purpose Despite widespread assumptions that negative social media content damages brand reputations, there is a lack of empirical data. This study aims to fill this research gap by determining if the volume or emotional intensity of actual electronic word-of-mouth (eWOM) content directed towards QSR brands on Twitter influences corporate reputation scores. Design/methodology/approach Using constructs from stakeholder and spiral of silence theories, a conceptual framework was first developed connecting eWOM and corporate reputation. A mixed-methods approach was used to test the model. More than 800,000 tweets directed towards eight QSR brands were collected over a 12-month period. Sentiment analyses quantified the valence, arousal, dominance, tabooness, offensiveness, and frequency of each word appearing in a branded conversation. Reputation Institute provided the reputation scores which were used as the dependent variable. Quantitative data was generated from a calculation of frequencies and percentages and used to statistically test hypotheses.Findings A strong, negative correlation (p=.027) between reputation scores and the average of high arousal scores (high_arousal_mean) was found. Therefore, eWOM measuring high in arousal is associated with decreasing reputation scores. Multiple regression analysis found that high arousal scores were only marginally significant in predicting QSR reputation scores (p=.058). Research limitations/implications There are several limitations to this study. First, only eight brands in a single service sector was examined. Second, the data only included tweets mentioning the brand by name, thereby possibly excluding tweets referencing the brand without explicitly mentioning it (e.g. Burger King as “BK”). Third, eWOM was only examined using the content posted to Twitter. Practical implications QSRs are told to monitor their brand mentions, but no one has yet told them which type of consumer eWOM would cause the most reputational damage. Realizing that high arousal scores can negatively influence corporate reputation scores for QSRs is an important step in helping managers deal with the volume of eWOM being created and shared on social media channels. Originality/value The study did not find empirical evidence linking angry tweets or electronic word-of-mouth (eWOM) high in negative valence to QSR reputation scores. However, the research did identify a significant correlation between the average of high arousal scores and corporate reputation.Keywords Corporate reputation, Electronic Word-of-Mouth, Sentiment Analysis, Valence-Arousal-Dominance, Social Media