Association for Public Policy Analysis & Management

Catastrophic natural disasters disrupt and change lives. These life changes can be temporary or permanent. Despite preparations and disaster relief programs provided either by the government or by the private household, disruptions and losses are inevitable. As a shared experience within a community, catastrophes create a sense of collective history among varied groups of citizens. Natural disasters produce public awareness that encourages the public’s evaluation of political actors, institutions, and policies. In this paper, we seek to uncover sentiments people have toward institutions and political actors in the event of a catastrophic natural disaster. We use the theoretical framework of affective attribution of blame developed by Atkeson and Maestas (2012), which argues that an extraordinary event triggers a need to seek accurate information without regard to predisposed positions which in turn evokes affective attribution of blame resulting in negative opinions about leaders and government.

We hypothesize that there are differences in the sentiments towards institutions or political actors by sex, geographical location, time of expressing sentiment relative to the time of the event, and level of a nation’s development. We use Big Data Analytics, particularly Twitter Analytics in R, to collect and analyze text data to measure differences in public sentiments towards the government or specific persons in the government during and after certain catastrophic events. To the best of our knowledge, there has been no study yet that uses Twitter Analytics to perform in-depth sentiment analysis for the attribution of blame in the event of catastrophic natural disasters. We will collect and analyze tweets around the time of impact of two catastrophic events from both the United States (a developed country) and the Philippines (a developing country). These countries both have a democratic government system and have approximately the same number of weather-related disasters for the period 1995-2015.

For the United States, we will be looking at Hurricane Matthew (2016) and Hurricane Irene (2011). These hurricanes are roughly similar with respect to fatalities and economic damages but different with respect to context (i.e., Hurricane Matthew occurred during an election period). For the Philippines, we will be looking at a cyclone event and a tectonic event occurring in the same year (2013). Both resulted in a large number of fatalities and economic damages. Our expected results will show that there are significant differences in public sentiments toward institution or specific persons across the dimensions of interest. More importantly, this study will present a way to use social media as an unsolicited feedback mechanism which political actors can use as part of the policy process to get evaluations of their performance in implementation of policies and programs.

References:
Atkeson, L. R., & Maestas, C. D. (2012). Catastrophic politics: How extraordinary events redefine perceptions of government. New York: Cambridge University Press.