Longitudinal Engagement and Audience Attention: Message Contents and Features that Get Passed On

Weather forecasts, alerts, and warnings, available through web and mobile applications such as Twitter, continue to reshape how most outside of the weather enterprise receive weather information. The effectiveness of how this information is communicated, and measurements of how individuals attend to it, has been an area of growing interest for both day-to-day and high-impact weather events. Identifying effective communication strategies allows practitioners to craft messages that engage individuals longitudinally and reach those who are at risk when they need information the most. In this presentation we will present our results from research investigating the longitudinal communication practices of 12 National Weather Service Weather Forecast Offices (WFOs), how those practices differ on calm and active weather days , and what specific practices attract public attention in good weather and bad.

Social media provide governmental organizations, such as the National Weather Service, the opportunity to communicate directly with stakeholders over long periods of time as a form of “online engagement.” Researchers have identified typologies of engagement that describe two types of content: message content, such as text that provides information about historical weather facts, forecasts, nowcasts, and weather products (advisories, alerts, and warnings); text that contributes to community building by inviting dialogue or interaction either online or in person; and text that inspires action or instructs individuals about what to do to prepare for or respond to hazardous weather (Olson et al. Under Review). A second type of content includes message “microstructural features,” which refer to the technological affordances of the social media channel that facilitate interaction and dialogue, including directed messages (@), hashtags (#), and URLs (Sutton et al., 2015). Recent research on longitudinal engagement practices of National Weather Service WFOs (Olson et al. Under Review) has found that offices share different message content and employ different microstructural features during calm (non-threat) weather periods compared to hazardous (threat) weather periods (indicated by the issuance of an advisory, watch, or warning). During calm weather, WFOs tend to share weather facts, engage in community building, and provide additional information via URL. During hazardous weather, they post more information about nowcasts, forecasts, and weather products while replying to more individual users.

Measuring engagement practices over time gives insight into communication strategies used by WFOs; however, little research has been conducted to identify the effect of these message strategies on behavioral outcomes (such as message passing among message receivers online), and whether those effects vary under different threat conditions. Previous research on message passing includes analyses of public safety organizations engaged in communication during periods of imminent threat and identified consistent message characteristics that increase the likelihood that they will be shared (Sutton et al. 2015; Vos et al. 2018), but little attention has been given to how messaging strategies affect message passing during nonthreat periods.

Our current research investigates tweets sent by twelve NWS WFOs in spring 2016. We use manual and automated coding procedures to identify engagement via message content and microstructural features (described above). We then conduct negative binomial regression modeling to identify features of messages that affect message passing in two time periods: threat and nonthreat. We find that the inclusion of actionable message content, information about historical weather facts, attached visual imagery (such as a map or infograph), and named event hashtags increases message passing during both threat and nonthreat periods. In contrast, messages that include forecast and nowcast information and messages that are sent in reply to other users have a lower passing rate. Findings suggest that common message features do affect message passing and that there are a set of message design strategies that have the potential for increasing the reach of engagement communication in varying time periods.

Olson, M. K., Sutton, J., Vos, S.C. , Prestley, R., Renshaw, S. L., & Butts, C. T. (under review). Build community before the storm: The National Weather Service’s use of social media. Journal of Contingencies and Crisis Management .

Sutton, J., Gibson, C. B., Phillips, N. E., Spiro, E. S., League, C., Johnson, B., Fitzhugh, S. M., & Butts, C. T. (2015). A cross-hazard analysis of Twitter warning messages. Proceedings of the National Academy of Sciences, 1112(48), 14793-14798.

Vos, S. C., Sutton, J., Yu, Y., Renshaw, S., Olson, M. K., Gibson, C. B., & Butts, C. T. (2018). Retweeting risk communication: The role of threat and efficacy. Risk Analysis.