The devil we know and the angel that did not fly: An examination of devil/angel shift in twitter fracking “debates” in NY 2008–2018

Abstract

Recent work has applied the Narrative Policy Framework (NPF) to examine narrative strategies in policy debates on social media platforms. We contribute to the literature by applying the NPF to fracking policy debates in New York using well‐established Natural Language Processing tools, including sentiment analysis. We combine this computational approach with a qualitative hand‐coding of pro‐ and antifracking Twitter influentials. This approach allows us to consider a much larger corpus of tweets over a much longer time frame than has been done thus far. We adapt and test NPF propositions related to the use of the devil/angel shift strategies before and after a major state‐wide policy change, that is, a state‐wide moratorium on high volume hydraulic fracturing or fracking. Overall, we find evidence for the use of the devil shift narrative strategy by the pro‐fracking coalition aimed at the Governor prior to the moratorium. After the moratorium, the relative percentage of Tweets containing devil shift sentiments decreases as the pro‐fracking coalition generally downshifts in its use of angel shift language without a corresponding increase in devil shift language, whereas, conversely, the anti‐fracking coalition generally downshifts in its use of devil shift language without a general increase in angel shift language. When we shifted our analysis to Tweets containing fracking and the Governor, we found a similar postban decrease in devil shift language among anti‐fracking users. Our findings offer lessons for using computational tools in the NPF as an approach to expand analytic ability and for the operationalization of concepts such as narrative strategies and policy entrepreneurs.

Shifting narrative strategies: How monument advocates change their stories in response to conflict over time

“…This supports Shanahan et al's (2013) observation that peaks in narrative construction centered around different points in the federal review process and were marred by the devil shift. While Pattison et al (2021) find a reduction in the devil shift following an event outside of advocates control, we believe this response was due to the fact that the event observed was conflict ending in that it issues a final moratorium on fracking in New York. As a conflict ending event, the observed decrease in the use of the devil shift by pro and anti-coalitions was likely part of a decrease in the use of all narrative strategies following the conclusion of a conflict.…”

“…This supports Sabatier et al's (1987) original development of the devil-angel shift within the Advocacy Coalition Framework and their assertion that it is used in times of high conflict. Thus, we assert the development of a hypothesis within the NPF that examines changes in the use of the devil-angel shift over the course of a policy conflict in response to key events such as those proposed by Pattison et al (2021). Although the identification of such key events will vary from case to case, the literature provides some areas for initial consideration including regulatory review, litigation filings as well as court rulings, and the issuance of executive orders.…”

“…In particular we evaluate the construction of narratives by organizations on Twitter. Our contribution to the NPF is, thus, twofold in that it (1) expands on initial work exploring the potential role of time in the construction of policy narratives by assessing variations in narrative strategies specifically and (2) examines these trends on social media which has been identified as an increasingly viable platform for the dissemination of policy narratives (Gupta et al, 2016; Merry, 2016; Pattison et al, 2021). Combined, we hope to further demonstrate the viability of temporal evaluations of policy narratives within the NPF and across media platforms.…”

The United States in Chinese environmental policy narratives: Is there a trump effect?

“…Relatedly, the database's inclusion and exclusion criteria, as well as whether and the extent to which the database itself engages in self‐censorship (e.g., avoid including articles of politically sensitive topics), seem less clear and warrant further scrutiny. These concerns may also apply to policy narratives/discourses drawn from other parts of the society, and call for finer‐grained operational strategies such as in‐depth interviews and social media analysis (see e.g., Ba, 2022; Merry, 2016; Pattison et al, 2022).…”

Political conflicts and surprising policy outcomes in times of crisis

“…In this contribution, too, online data, especially politicians' websites on the one hand and critical blogs on the other hand, form the most important data basis (in Schlaufer et al's case supplemented by interviews). As a result, the recently popularized concepts of Angel Shift and Devil Shift (Pattison et al, 2021;Stephan, 2020) enable the identification of different narrative strategies of the governmental coalition on the one hand and oppositional actors on the other.…”