SWEAT: Scoring Polarization of Topics across Different Corpora

TL;DR

This paper introduces SWEAT, a new statistical method for measuring the polarization of topics across different text corpora, aiding social science research.

Contribution

The paper presents SWEAT, a novel measure that quantifies topic polarization using distributional representations and opposite-valence wordsets.

Findings

SWEAT effectively measures polarization differences.

Validation confirms SWEAT's reliability.

Case study demonstrates practical utility.

Abstract

Understanding differences of viewpoints across corpora is a fundamental task for computational social sciences. In this paper, we propose the Sliced Word Embedding Association Test (SWEAT), a novel statistical measure to compute the relative polarization of a topical wordset across two distributional representations. To this end, SWEAT uses two additional wordsets, deemed to have opposite valence, to represent two different poles. We validate our approach and illustrate a case study to show the usefulness of the introduced measure.