Association for Public Policy Analysis & Management

"How do focusing events affect policy narratives? Do narratives vary in time and space from focusing events? In this paper, we investigate the role of focusing events on two types of policy conflicts related to urbanization of bats in Australia. The first is the mass mortality of bats due to sudden changes in temperature and humidity. We investigate the effect of mass mortality events on narratives related to nuisance and flying fox management in urban areas. The second is the spillover of deadly zoonotic diseases from bats to horses to humans. We investigate the effect of disease spillover on narratives related to public health. Narratives are a vehicle for communication built on a universal structure that describes causal relationships between events in a setting that affect a set of characters (Shanahan 2013). The policy realities constructed by narratives are perceived as truths, sometimes in the face of countervailing evidence (Shanahan 2011). We hypothesize that focusing events, such as mass bat mortalities or disease spillovers, crystallize policy narratives, intensify emotional responses, and polarize the dominant stakeholder narratives. To measure narratives across space and time we examine newspaper media reporting before and after focusing events. We collect content from a selection of national, regional, and local newspapers from 1990-2018. We begin before 1994 in order to evaluate reporting before the discovery of lyssavirus and Hendra. We conduct three types of text analysis on our corpus of newspaper content. First, we conduct a sentiment analysis on all newspaper content. Sentiment analysis determines the overall tone of each article by calculating the difference between the number of positive words and the number of and negative words in the article (Hu and Liu 2004). We construct a baseline sentiment for each newspaper source and compare the baseline sentiment to sentiment captured in articles explicitly mentioning key words associated with the policy issue (human-bat conflict). We then overlay timing of the focusing events to evaluate the relationship between focusing events and changes in sentiment. Second, we isolate the sub-corpus of reporting specifically relevant to human-bat conflict. In order to identify changes in perceived villains and victims in the human-bat conflict, we tag parts of speech and isolate nouns as potential villains and victims. We then categorize them according to the positive or negative valence of the surrounding words. We investigate to what extent the population of villains and victims changes as distance from focusing events increases in time and geographic space. Finally, we isolate a smaller sample of newspaper articles at varying distances from focusing events and map the arc of their narratives using Plot Mapper (Beauchamp, paper). This descriptive analysis and visualization serves to highlight changes in the temporal structure of narratives as distance from a focusing event increases."