Cutting Through the Noise: Predictors of Successful Online Message Retransmission in the First 8 Months of the COVID-19 Pandemic

Abstract

In this paper, we investigate how message construction, style, content, and the textual content of embedded images impacted message retransmission over the course of the first 8 months of the coronavirus disease 2019 (COVID-19) pandemic in the United States. We analyzed a census of public communications (n = 372,466) from 704 public health agencies, state and local emergency management agencies, and elected officials posted on Twitter between January 1 and August 31, 2020, measuring message retransmission via the number of retweets (ie, a message passed on by others), an important indicator of engagement and reach. To assess content, we extended a lexicon developed from the early months of the pandemic to identify key concepts within messages, employing it to analyze both the textual content of messages themselves as well as text included within embedded images (n = 233,877), which was extracted via optical character recognition. Finally, we modelled the message retransmission process using a negative binomial regression, which allowed us to quantify the extent to which particular message features amplify or suppress retransmission, net of controls related to timing and properties of the sending account. In addition to identifying other predictors of retransmission, we show that the impact of images is strongly driven by content, with textual information in messages and embedded images operating in similar ways. We offer potential recommendations for crafting and deploying social media messages that can “cut through the noise” of an infodemic.

The social amplification and attenuation of COVID-19 risk perception shaping mask wearing behavior: A longitudinal twitter analysis

“…Our study describes six ways in which COVID-19 risk was discussed in mask tweets on social media, a platform that functions as a risk amplification and attenuation station through the nature of the content but also through mentions to government officials, journalists, and politician stakeholders (e.g., @POTUS, @Trump, @governor) [ 65 ]. SARF contends that social amplification of risk occurs at two stages: in the initial transmission of information about a risk (we examined the social media context) and in the response mechanisms of society (behavioral, economic, and symbolic impacts).…”

A Framework of Social Media Messages for Crisis and Risk Communication: A Study of the Covid-19 Pandemic

“…For example, a popular "genre analysis" proposed 5 "top-level genres" (broadcast information, broadcast warning, encourage behavior, appeal for information, fighting rumours) to categorize flood [13] and earthquake emergency communications [12]. Others have examined "condolences" and "encouragement" [4] messages, while recent studies examined "resilience" and "susceptibility" content, among others [10,11].…”

“…as expressives [25]. It is important to show empathy in crisis communication [1], and others have discussed the use of "emotion-evaluative" [3], "resilience" [11] and "reassurance" message features in related contexts [9]. The presence of "emotion-evaluative" content has been found associated with higher message diffusion in multiple CRC scenarios [3], while "resilience" keywords had a more mixed but positive relationship in the context of Covid-19 [11].…”

“…Despite the importance of a social media presence, there is no unifying framework for analysis of government social media messages during crises. Existing frameworks seem to use ad-hoc [8,9], or a purely based on a lexical approach [10,11]. Studies often do not distinguish intention of messages (e.g.…”

“…Previous studies have distinct purposes and have contributed to our understanding of crisis and risk communication messages on social media. However, the "categories", [9,15], "strategies" [16], "frames" [9] or "features" [11] of crisis and risk messages could benefit from a more linguistically informed framework; one that distinguishes speech acts (e.g. message purpose or intention) from topics (e.g.…”