Moral and emotional influences on attitude stability towards COVID-19 vaccines on social media

Abstract: Effective public health messaging benefits from understanding antecedents to unstable attitudes that are more likely to be influenced. This work investigates the relationship between moral and emotional bases for attitudes towards COVID-19 vaccines and variance in stance. Evaluating nearly 1 million X users over a two month period, we find that emotional language in tweets about COVID-19 vaccines is largely associated with more variation in stance of the posting user, except anger and surprise. The strength of COVID-19 vaccine attitudes associated with moral values varies across foundations. Most notably, liberty is consistently used by users with no or less variation in stance, while fairness and sanctity are used by users with more variation. Our work has implications for designing constructive pro-vaccine messaging and identifying receptive audiences.

• The paper reveals that anti-vaccine tweets use significantly more moral and emotional cues, particularly in themes of sanctity, liberty, fear, and sadness.
• It employs a lexicon-based analysis and BERTweet stance detection across nearly one million vaccine-related tweets collected from March to May 2021.
• The findings link specific emotional expressions with attitude stability, offering insights to design targeted public health interventions.

Analysis of Moral and Emotional Language in COVID-19 Vaccine Discussions on Social Media

The paper "Moral and emotional influences on attitude stability towards COVID-19 vaccines on social media" by Phillips et al. undertakes an insightful exploration of how moral and emotional language expressed in social media posts influences the stability of attitudes towards COVID-19 vaccines. By conducting an extensive analysis of tweets from the platform X (previously known as Twitter) from March to May 2021, the researchers aimed to examine how moral and emotional cues relate to changes in stance regarding COVID-19 vaccines.

The study focuses on three primary research questions (RQs):

1. How is moral and emotional language utilized differently in pro- and anti-COVID-19 vaccine tweets?
2. How does the use of this language differ between users with a stable stance and those with variable stances?
3. To what extent does the inclusion of moral and emotional language correlate with variations in stance over time?

Methodology

Using a lexicon-based approach, the authors analyzed nearly 1 million users who tweeted about COVID-19 vaccines, ultimately filtering tweets related to vaccines through various keywords. A stance detection classifier based on BERTweet was developed to tag each tweet as pro, anti, or neutral regarding vaccines. The aim was to correlate these stances with the presence of moral foundations and emotional expressions identified through Netmapper software.

The stance variance, a proxy measure for attitude strength, was calculated using the standard deviation of stances expressed by users across tweets. The authors then analyzed the associations between stance variability and the use of moral values—care, fairness, liberty, sanctity, loyalty, authority—and emotions—happiness, sadness, fear, anger, disgust, surprise.

Main Findings

The analysis revealed that anti-vaccine tweets incorporate more moral and emotional language compared to pro-vaccine tweets, with references to sanctity, liberty, and emotions such as fear and sadness being prevalent. On the contrary, pro-vaccine tweets showcased more loyalty and happiness.

Exploring RQ2, users with a consistent stance employed emotional language like anger, and moral foundations like loyalty and liberty more frequently than those with variable stances. Conversely, users with varying stances leaned towards expressions of emotions like happiness and surprise alongside moral foundations like fairness and sanctity.

In addressing RQ3, the study identified that emotions generally contributed to greater variation in stance, with exceptions like surprise—which is more associated with less outlier-prone stance fluctuation—and anger, where the effect was not statistically significant.

Discussion and Implications

This research contributes meaningfully to understanding the interplay between moral convictions, emotional expressions, and attitude stability concerning public health topics on social media. While past studies have touched upon moral values' role in shaping vaccine hesitancy, this work provides empirical evidence on how these factors influence the stability of vaccine-related attitudes over time.

Practically, these findings have implications for designing effective public health messages and interventions. Messaging that recognizes the emotional and moral language associated with less stable attitudes could be more effective in persuading or informing audiences on vaccine-related issues, a vital aspect amidst the ongoing battle against misinformation.

Future Directions and Limitations

Future research might expand to other social media platforms, providing a broader scope on how moral and emotional expressions influence public health attitudes in different digital environments. Additionally, analyzing the directionality of stance changes can sharpen understanding, enabling targeted health communication strategies to reinforce positive behavioral shifts.

While this study offers extensive insights, it is limited by its focus on a single platform and by not accounting for automated bot behavior that might skew data assessments. Addressing these limitations could refine future analyses and provide more robust guidelines for public health communication strategies in the social media age.