Does Every Cloud Have a Silver Lining? Disasters and Polarization

Abstract

How do natural disasters impact political polarization? Which cues and framings work for de-polarizing the electorate during natural disasters? This study aims to conduct a sentiment analysis of social media data to examine the impact of natural disasters on polarization and then several survey experiments in Turkey. This highly polarized country recently went through natural disasters, including floods and wildfires. By randomizing the partisan identity of donor organizations (NGOs, municipalities) and the aid recipients, this project will help us understand to what extent the opposition’s de-polarization strategies are effective under an electoral autocratic and populist regime.

Background

This study aims to examine the effectiveness of opposition party solidarity strategies on depolarization under electoral-populist autocracies. First, we will conduct a sentiment analysis of social media data to examine the impact of natural disasters on polarization and then several survey experiments in Turkey. This highly polarized country recently went through natural disasters, including floods and wildfires. By randomizing the partisan identity of donor municipalities and the aid recipients, this project will help us understand to what extent the opposition’s de-polarization strategies are effective under an electoral autocratic and populist regime.

Research Design

In the first between-subject experimental design, we aim to measure the impact of (1) significant cross-partisan disaster relief aid by the opposition party municipalities in pro-government or pro-Kurdish party districts post-disaster, and 2) the effect of government counter-propaganda against such disaster relief programs on respondent’s level of affective polarization.

Each respondent will be assigned to either a control group or one of three treatment conditions.

(1) The control condition presents a non-fictitious narrative that indicates how disaster aid was poured into several cities after natural disasters with no partisan signaling or reference to any actor’s name.
(2) The first treatment consists of two layers: In addition to the control narrative, an extra non-fictitious text elaborates how specific opposition municipalities offered significant financial assistance to pro-government cities post-disaster by adopting a cross-partisan discourse. The text signals the opposition political elite’s willingness to communicate across the aisle.
(3) The second treatment consists of three layers: In addition to the control AND first treatment narrative, the added text includes a pro-government news reporting that tries to discredit the opposition party’s disaster relief programs.
(4) The third treatment presents a non-fictitious narrative that includes the control narrative AND an added text that shows how opposition municipalities offered help to pro-Kurdish party districts post-disasters. There are two treatments for pro-government and pro-Kurdish party districts because polarization is not bi-partisan in Turkey. In addition to the high level of polarization between pro-Cumhur (incumbent alliance) and pro-Millet (main opposition alliance) voters, pro-HDP voters are cited as the “most distant” group in nationally-representative surveys.

In the second between-subject experimental design, we aim to assess the effect of ultra-nationalist/militarist fiction/TV series on public support for autocratic repression and cross-border military operations. Each respondent will be assigned to either a control group or one of the treatment conditions. (1) The control group will be allowed to choose one clip from a list of four TV series. These clips are selected from non-political Turkish dramas broadcasted by the government TV station TRT. (2) In the first treatment condition, similarly, we will provide four options for the respondent to choose. All four clips will be from militarist TV series broadcasted by either TRT or pro-government TV stations. (3) All four clips will be retrieved from TRT’s historical-Ottomanist genre in the second treatment condition, and the respondent will choose one.

We will use block randomization, where participants will be stratified based on their party identification (that is, whether they are supporters of the incumbent electoral alliance, the main opposition alliance, or the pro-Kurdish party), and then randomly assigned to treatment conditions.

Hypotheses

• H1.1: Controlling for other factors, we hypothesize that government supporters, who are exposed to news about opposition municipalities’ significant aid to pro-government districts in the aftermath of a natural disaster, on average, will show a more favorable attitude towards opposition party supporters.
• H1.2: This positive impact will be mitigated by the government’s propaganda against the opposition party’s disaster relief programs but will remain significant.
• H2.1: Controlling for other factors, we hypothesize that opposition party supporters, who are exposed to news about opposition municipalities’ significant aid to pro-government districts in the aftermath of a natural disaster, on average, will show a more favorable attitude towards government supporters.
• H2.2: This positive impact will be mitigated by the government’s propaganda against the opposition party’s disaster relief programs but will remain significant.
• H3.1: Pro-Kurdish party voters, who are exposed to news about opposition municipalities’ significant aid in predominantly Kurdish regions following a natural disaster, on average, will show a more favorable attitude towards opposition party supporters.
• H3.2: Pro-government voters, who are exposed to news about opposition municipalities’ significant aid in predominantly Kurdish regions following a natural disaster, on average, will show a less favorable attitude towards opposition party supporters.