JMIR Publications

ABSTRACT

Background:

Public attitudes toward health issues are becoming increasingly polarized, as evident in social media comments that range from supportive to oppositional and often contain uncivil language. The combined effects of comment slant and comment tone on health behavior among polarized publics need further examination.

Objective:

We aim to examine how social media users’ prior attitudes toward mask-wearing and their exposure to a mask-promoting post synchronized with polarized and hostile discussions affect their compliance with mask-wearing.

Methods:

The study was a web-based experiment with participants recruited from Amazon Mechanical Turk. A total of 522 participants provided consent and completed the study. Participants were automatically and randomly assigned to be exposed to a mask-promoting post accompanied by either civil anti-mask-wearing comments (n = 130), civil pro-mask-wearing comments (n = 129), uncivil anti-mask-wearing comments (n = 131), or uncivil pro-mask-wearing comments (n = 132), and then completed self-assessed questionnaires. The PROCESS Macro in SPSS (Model 12) was employed to test the three-way interaction effects between comment slant, comment tone, and prior attitudes on participants’ presumed influence of the post and their behavioral intention to comply with mask-wearing.

Results:

Anti-mask-wearing comments led social media users to presume less influence about others’ acceptance of masks (B = 1.49, P < .001, 95% CI 0.98-2.00) and resulted in decreased mask-wearing intention (B = .07, P = .03, 95% CI 0.01-0.13). Comment tone with incivility, also, reduced compliance with mask-wearing (B = -.44, P = .02, 95% CI -0.81 to -0.07). Furthermore, polarized attitudes not only had a direct influence (B = .86, P < .001, 95% CI 0.45-1.26), but also interacted with comment slant and comment tone, influencing mask-wearing intention (B = -.84, P = .03, 95% CI -1.59 to -0.09).

Conclusions:

Pro-mask-wearing comments enhanced presumed influence and compliance of mask-wearing, but incivility in the comments hindered this positive impact. Anti-maskers showed increased compliance when they were unable to find civil support for their opinion in the social media environment. The findings suggest the need to correct and moderate uncivil language and misleading information in online comment sections while encouraging the posting of supportive and civil comments. Additionally, information literacy programs are needed to prevent the public from being misled by polarized comments.