What effects should politicians expect by employing a particular communicative behaviour? Are measures of sentiment and emotion complimentary or interchangeable in understanding the reception on social media? Much extant research implies particularly that – stemming from the theory of negativity bias – the use of negative sentiment should be an effective strategy. Diverging effectiveness in the use of negative language begs the question as to whether more nuanced measures such as emotion could provide a clearer picture. This study tests these theories implementing a micro-analysis of UK politicians’ Twitter/X communications (N=719,603). My findings reveal that, while neutral messages predominate, positive sentiments can resonate well with the public, contradicting established beliefs. The research, controlling for a range of fixed effects, reveals also a picture more nuanced than the frequently used sentiment-based approach. Language employing emotions like joy and disgust influence the popularity of tweets strongly whilst by no means always combining with positive and negative sentiment as expected respectively. Disgust leads to the highest approval and fear-based messages find least favour. Different implied emotions have varying impacts on sentiment, both in magnitude and polarity.
Sentiment is a broader evaluative measure that is shaped by a combination of factors, including but not limited to emotion. This would motivate a new emphasis on treating sentiment and emotion as complementary but distinct components in social media analysis. It is, however, only with the quantification of emotion-types at scale that these results are revealed.