Looking beyond the Crisis of Democracy: Patterns of Representation in Israeli Elections.
Western democracies have been suffering for several decades from the “crisis of representative democracy”. This crisis is evident in the citizenry’s disaffection with politics and politicians, declining voter turnout, and the upsurge of antiestablishment and populist candidates and parties, among other phenomena. Against this backdrop we ask whether we witness the end of representative democracy or rather a vital change in the manifestations of representative democracy. Based on Pitkin’s classic multidimensional conception of political representation, our project offers a harmonious analysis of these dimensions together in the Israeli political context, as pronounced in the discourse of (1) three major actors in democracies (politicians, media, and citizens); (2) across platforms (speeches, policies, social media, news coverage, and public opinion surveys); (3) while taking into account the reciprocal relationships among these actors and platforms; and (4) and the longitudinal dynamics. By using our state-of-the-art computational algorithm for content analyses of big data, we hypothesize that politicians represent their citizens in much more symbolic ways (by emphasizing shared experiences, narratives, and values) than substantive (actual policy activity). This is due to the increasing personalization and populism in Israel, the rise in collective identity considerations in Israeli politics, and the rise of social media—which enables representatives to address their constituents directly. Moreover, by gauging a wide range of elements in the Israeli political discourse—values, narratives, frames, issues, and vocabulary—we will examine whether changes in the discourse regarding representation result from developments among the public or from strategic manipulations by political elites.
This project is undertaken as part of both the Israel National Election Studies with funding from the Israel Science Foundation (ISF), as well as with a support from the Social Science Research Council’s Social Media and Democracy Research Grants.
We are developing advanced machine learning and deep learning technologies for highly accurate text analysis. Combining narrative approaches from the social sciences with automatic textual analysis, we are building a novel system that detects meaningful discourse trends, including stories associated with industries and companies, providing accurate insights for users.
The “Davis index” aims to represent international media attention on the Israeli-Arab conflict and relations, focusing on the Israeli-Palestinian conflict.
The index aggregates a list of general categories, such as “Middle East Conflicts” and “Israel and the Arab world”; as well as specific categories, such as “Palestinian organizations”, “the BDS movement”, and interventions by specific governments or international institutions. The index is constructed through automated textual analysis, through a process combining topic modelling, deep learning and expert coding. The method incorporates topic models for its unsupervised component and deep learning for its supervised component. Both stages take the context into account, while focusing on the sentence as the desired unit of analysis. The result is a multi-label text classification method - attributing multiple topics to each sentence.
Throughout the 2018-19 year, we presented our results for the Davis Index Reports, publishing the reports together with the Leonard Davis Institute for International Relations. We have produced the first report for the year 2010, and plan to release reports for the years 2011-2018 and onward. Our plan is to publish an updated report every three months, and an overall yearly report at the end of each calendar year.
Narrative Chambers and News Sharing on Facebook
This project analyzes the content of news articles viewed and shared on Facebook. It is designed to uncover the active role played by users in the competition over the dissemination of social narratives. The project is supported by the Social Science One Industry-Academic Partnership (https://socialscience.one/).