TL;DR
This study reveals that misinformation about e-cigarette harms on Twitter influences adult smokers' purchase intentions not just through perceived risks, but also through a complex interplay of discrete emotions. Specifically, emotions like worry, hope, and happiness mediate these effects, and in some cases, emotions sequentially influence perceived harm before purchase intent, highlighting the critical role of affect in navigating health misinformation.
What Problem Does This Paper Address?
The proliferation of online misinformation, particularly concerning e-cigarettes, poses a significant public health challenge. While we know misinformation can influence behavior, the underlying psychological mechanisms, especially the role of emotions, are not well understood. This paper addresses this gap by investigating how exposure to misinformation about e-cigarette harms on Twitter affects adult smokers' intentions to purchase e-cigarettes, focusing specifically on how discrete emotional responses and perceived relative harm mediate this relationship. Understanding these affective pathways is crucial for designing effective interventions to counter health misinformation.
Methods
The researchers employed a web-based randomized controlled experiment. They recruited 2400 adult smokers and randomly assigned them to view one of four conditions of tweets:
- E-cigarettes are as or more harmful than smoking (misinformation).
- E-cigarettes are completely harmless (misinformation).
- Uncertainty of e-cigarette harm (misinformation).
- Control tweets about physical activity.
After exposure, participants completed post-test measures of emotional responses, perceived relative harm of e-cigarettes, and intention to purchase e-cigarettes. They then used structural equation modeling with bootstrapping to analyze mediation pathways, examining direct and indirect effects through perceived relative harm and six discrete emotions (scared, angry, worried, happy, hopeful, relieved).
Datasets
The study utilized data collected from a web-based experiment involving 2400 adult smokers recruited from the US and UK. The "dataset" in this context is the experimental data generated from participant responses to the manipulated tweet conditions and subsequent questionnaires. No pre-existing datasets were used for training or evaluation; rather, the study generated its own data through a controlled experiment.
Key Insights and Ideas
- Emotions are Key Mediators: It's not just about facts (or misinformation); how people feel about the information is a critical driver of their behavior. This moves beyond purely cognitive models of misinformation effects.
- Discrete Emotions Matter: The study highlights that not all negative emotions have the same impact. Specific emotions like "worried," "hopeful," and "happy" were significant mediators, suggesting targeted emotional appeals in interventions could be more effective.
- Sequential Emotional Influence: The finding that emotions can serially mediate the effect (emotion -> perceived harm -> purchase intention) is particularly interesting. This suggests a dynamic process where initial emotional reactions shape risk perceptions, which then influence behavior.
- Affective and Cognitive Interplay: The research provides preliminary evidence that emotions and cognitive assessments (perceived harm) are deeply intertwined in how misinformation is processed and influences behavior.
Key Results
- Perceived Relative Harm Mediation: Exposure to misinformation in Condition 1 (more harmful) led to lower purchase intentions, partly mediated by increased perceived relative harm. Conversely, Condition 2 (completely harmless) led to higher purchase intentions due to lower perceived relative harm.
- Direct Emotional Mediation: Three emotions significantly mediated the effect of misinformation on purchase intention:
- Worried, Hopeful, and Happy: Mediated the effects of Condition 1 vs. Control.
- Hopeful and Happy: Mediated the effects of Condition 2 vs. Control.
- Worried, Hopeful, and Happy: Mediated the effects of Condition 3 vs. Control.
- Serial Mediation: Nine out of 18 serial mediation paths through emotions and perceived relative harm were significant. Specifically, emotions like Scared, Worried, Angry, and Hopeful played a role in this sequential pathway.
- No US/UK Differences: The study found no significant differences in mediation effects between US and UK participants.
Limitations and Future Work
- Self-Reported Measures: The study relied on self-reported emotional responses, perceived harm, and purchase intentions. Future research could incorporate psychophysiological measures (e.g., skin conductance, heart rate) for more objective emotional arousal assessment.
- Causal Direction: The cross-sectional nature of post-test measures makes it difficult to definitively establish the causal direction between emotions and perceived relative harm. While sensitivity analyses were conducted, further research is needed to untangle these causal relationships.
- Brief Exposure: The study involved a brief, single exposure to misinformation. Future work should investigate the effects of repeated exposure in more naturalistic social media settings.
- Population Subgroups: The study did not examine whether mediating pathways differ across various population subgroups. Future research should explore how factors like vulnerability to misinformation (e.g., conspiracy ideation, health literacy) might influence these relationships.
Summary updated