Climate change believers and sceptics: a text-mining approach to measure their impact on the political discourse

点击翻译按钮获取中文摘要

A growing interdisciplinary literature examines the role of industry actors and key conservative think tanks for shaping the climate change debate and promoting inaction. It explores how the climate change counter movement (CCCM) tries to undermine legislative attempts of restricting carbon emissions by manufacturing doubt about the credibility of anthropogenic climate change.The aim of my research is to quantify the influence of CCCM on the political discourse with a text-mining approach. Prior attempts to quantify the impact of the CCCM on political discussion, using a simple measure of semantic similarity, have found a positive relationship over time between the discourse of the CCCM with both the news media and the US presidents, i.e. their discourse is becoming more similar, thus indicating an increasing influence of the CCCM. However, the measure of semantic similarity used does not isolate words that are (at least on average) unique to the CCCM and thus may be overstating the "influence" of these actors on political speech. As such, the positive relationship is possibly due to an overall increase in climate change discourse and not necessarily as a result of the CCCM.The goal of this research is, therefore, to isolate the sceptical language in a first step, by taking the competing discourse of the environmental movement into account. The environmental movement and the CCCM are thought to be contesting the definition of the dominant field frame in the climate change debate forming diametrically opposed movements. Distinguishing between the two discourses will allow a more nuanced exploration of political agendas, while also providing a suitable measure of the impact of each group. Specifically, the following questions should be explored:1) What are the best methods for isolating the language of both sceptics and the convinced? Working closely with my advisor, I will explore a wide range of supervised and semi-supervised methods to classify movement language.2) How can one measure "influence" in political speech?3) Is there a measurable impact of the climate change discourse(s) on the political discourse and can the origin of this impact be traced?