Addressing Antivaccine Sentiment on Public Social Media Forums Through Web-Based Conversations Based on Motivational Interviewing Techniques: Observational Study
Bibliographic record
Abstract
BACKGROUND: Health misinformation shared on social media can have negative health consequences; yet, there is a dearth of field research testing interventions to address health misinformation in real time, digitally, and in situ on social media. OBJECTIVE: We describe a field study of a pilot program of "infodemiologists" trained with evidence-informed intervention techniques heavily influenced by principles of motivational interviewing. Here we provide a detailed description of the nature of infodemiologists' interventions on posts sharing misinformation about COVID-19 vaccines, present an initial evaluation framework for such field research, and use available engagement metrics to quantify the impact of these in-group messengers on the web-based threads on which they are intervening. METHODS: We monitored Facebook (Meta Platforms, Inc) profiles of news organizations marketing to 3 geographic regions (Newark, New Jersey; Chicago, Illinois; and central Texas). Between December 2020 and April 2021, infodemiologists intervened in 145 Facebook news posts that generated comments containing either false or misleading information about vaccines or overt antivaccine sentiment. Engagement (emojis plus replies) data were collected on Facebook news posts, the initial comment containing misinformation (level 1 comment), and the infodemiologist's reply (level 2 reply comment). A comparison-group evaluation design was used, with numbers of replies, emoji reactions, and engagements for level 1 comments compared with the median metrics of matched comments using the Wilcoxon signed rank test. Level 2 reply comments (intervention) were also benchmarked against the corresponding metric of matched reply comments (control) using the Wilcoxon signed rank test (paired at the level 1 comment level). Infodemiologists' level 2 reply comments (intervention) and matched reply comments (control) were further compared using 3 Poisson regression models. RESULTS: In total, 145 interventions were conducted on 132 Facebook news posts. The level 1 comments received a median of 3 replies, 3 reactions, and 7 engagements. The matched comments received a median of 1.5 (median of IQRs 3.75) engagements. Infodemiologists made 322 level 2 reply comments, precipitating 189 emoji reactions and a median of 0.5 (median of IQRs IQR 0) engagements. The matched reply comments received a median of 1 (median of IQRs 2.5) engagement. Compared to matched comments, level 1 comments received more replies, emoji reactions, and engagements. Compared to matched reply comments, level 2 reply comments received fewer and narrower ranges of replies, reactions, and engagements, except for the median comparison for replies. CONCLUSIONS: Overall, empathy-first communication strategies based on motivational interviewing garnered less engagement relative to matched controls. One possible explanation is that our interventions quieted contentious, misinformation-laden threads about vaccines on social media. This work reinforces research on accuracy nudges and cyberbullying interventions that also reduce engagement. More research leveraging field studies of real-time interventions is needed, yet data transparency by technology platforms will be essential to facilitate such experiments.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".