US-IALE 2018

AUTHORS: Alexander Killion*, Boise State University; Neil Carter, Boise State University; Tracy Swem, Michigan State University

ABSTRACT: Human-caused mortality of carnivores has significantly increased the extinction risk for many of the world’s large carnivore species. On the other hand, large carnivores provide a variety of benefits to human societies. However, the risks and benefits from carnivores are often unequally distributed in human landscapes. This asymmetry has contributed to a polarizing public discourse about their conservation and has made it difficult to develop non-controversial legislation. Since collecting social survey data to examine temporal trends in public opinion is costly and time-consuming, analyzing the salience of media reports is an efficient alternative. Public opinion influences what the media reports, and by disseminating certain types of information, the media has in turn been shown to influence human perception of wildlife risk. However, there are often computational constraints in analyzing a large corpus with sufficient detail to capture trends in salience over a long timeframe that has witnessed changes in both policies and carnivore abundances. To overcome these constraints, we utilized a hierarchical Bayesian framework to identify distinct topics of highest salience regarding gray wolves (Canis lupus) and assess trends in the prevalence of those topics over 55 years (1960-2015), the longest wildlife media content analysis to date. We highlight significant differences between local and national salience levels over time and present how our methodology can identify changes in public opinion in heterogeneous social landscapes.