Topics and Sentiments Influence Likes: A Study of Facebook Public Pages' Posts About COVID-19 Vaccination.
Cyberpsychol Behav Soc Netw
; 25(9): 552-560, 2022 Sep.
Article
in English
| MEDLINE | ID: covidwho-1992054
ABSTRACT
In this study, we analyzed >200,000 posts collected from Facebook public pages that were published in December 2020 during the rollout of the first dose of the Pfizer-BioNTech vaccine to the American public. We ran both topic modeling and sentiment analysis of the posts and found that first Facebook posts talked about not only treatment effectiveness such as trial results and testing approaches but also other issues that surround vaccines such as approval and distribution. Second, although the general sentiment during this period was positive and anticipation was the highest emotion, Facebook posts expressed a salient amount of fear and sadness, especially when discussing emergency approval and trial results. Finally, we found that both topics and sentiments have a significant influence on user likes. Using the topic of vaccine distribution as a baseline, posts that discuss related aspects of vaccines (e.g., effectiveness, shipment, and testing), call for actions (e.g., use of masks) and indicate care to vulnerable groups (e.g., health care workers and seniors) received more likes. Posts with bad news (e.g., new cases and deaths) and doubts over the usefulness of vaccines received fewer likes. Regardless of their valences, approach emotions lead to more likes whereas withdrawal emotions lead to fewer likes. Our study suggests that to facilitate actions, using certain topics and approach emotions in the posts could be helpful.
Keywords
Full text:
Available
Collection:
International databases
Database:
MEDLINE
Main subject:
Social Media
/
COVID-19
Type of study:
Experimental Studies
/
Observational study
/
Randomized controlled trials
Topics:
Vaccines
Limits:
Humans
Language:
English
Journal:
Cyberpsychol Behav Soc Netw
Journal subject:
Behavioral Sciences
/
Psychology
Year:
2022
Document Type:
Article
Affiliation country:
Cyber.2022.0063
Similar
MEDLINE
...
LILACS
LIS