When Do People Verify and Share Health Rumors on Social Media? The Effects of Message Importance, Health Anxiety, and Health Literacy

Abstract

This study explores the roles of perceived message importance, health anxiety, and health literacy in the relationship between message factors (message label and message valence) and behavioral intentions for rumor verification and sharing. 660 Twitter users responded to unverified information regarding the influenza vaccine. A 3 (label: none vs. news vs. rumor) × 2 (valence: positive vs. negative) online semi-experiment, with a survey to measure health anxiety and health literacy, showed the following results: First, perceived message importance mediated the relationship between message factors and behavioral intentions: only in the condition of the negative message, participants considered a news-labeled message more important than a rumor–labeled or a no-label message. Perceived message importance was associated with intentions to verify and share the message. Second, health anxiety interacted with perceived message importance only when predicting an intention to share the message. Last, healthy literacy interacted with perceived message importance when predicting intentions to both verify and share the message. The results will provide implications for health communication research and practices, especially on managing and controlling rumor dissemination on social media.

Notes

¹ Vosoughi et al. (2018) examined more than 126,000 stories on twitter, and found that, on average, rumors spread to 1,500 users in 10 hours while verified news needed 60 hours to reach the same number of people.

² Retrieved from 2019 Global Digital Overview (https://wearesocial.com/global-digital-report-2019).

³ For the pretest, 50 people were recruited through the same means as the participants for the main study: they were from the panel of an online research company, and their demographics (e.g., sex, age) shadowed the proportion of the Korean population.

⁴ Andrew Hayes’ model 8 tests the moderated mediation effect, where an independent variable (x) and a moderator (w) jointly influence a mediator (m) and a dependent variable. It also tests whether the effect of the two predictors (x, w) on the dependent variable is mediated by the mediator (m). Message valence was treated as an independent variable (x), and message label was treated as a moderating variable (w). Perceived message importance was a mediator (m) between the two predictors (x, w) and the dependent variable (y), which was either an intention to verify the message (y₁) or share the message (y₂).

⁵ Andrew Hayes’ moderated mediation model 61 tests the effects of two moderators and one mediator. It tests if an independent variable (x) and a moderator (z) jointly influence a mediator (m), and whether an independent variable (x) and another moderator (w) jointly and directly influence a dependent variable (y). It also tests if the mediator (m) mediates between the relationship between the three predictors (x, z, w) and the dependent variable (y), and whether the mediator (m) interacts with another moderator (w) when it influences the dependent variable (y). Message valence was entered as an independent variable (x) and message label was entered as a moderating variable (z) between valence and perceived message importance (m). Perceived importance was a mediator (m) between the two predictors (x, w) and the dependent variable (y), either intention to verify (y₁) or share the message (y₂). Health anxiety was entered as another moderator (w₁) for Research Question 2 and health literacy was entered as a moderator for Research Question 3 (w₂).

⁶ We also examined the interaction between message valence and label using 2-Way ANOVAs. There was a significant interaction between valence and label in predicting perceived message importance (F (2, 654) = 10.06, p < .001, partial ƞ² = .03). A post-hoc analysis indicated that only in the condition of the negative message, a news-labeled message was perceived as more important (M = 3.10, SD = .90) than a rumor–labeled message (M = 2.55, SD = 1.00) or the no-label message (M = 2.47, SD = 1.06) (p < .001).