Agreeing to Disagree: Annotating Offensive Language Datasets with Annotators' Disagreement

TL;DR

This paper investigates how annotator disagreement impacts offensive language detection datasets and classifiers, highlighting the importance of including ambiguous cases to improve model robustness and better reflect online subjectivity.

Contribution

It introduces three new datasets with varying levels of annotator agreement and demonstrates the significant effect of disagreement on classifier performance and robustness.

Findings

Selecting data based on agreement levels affects classifier accuracy.

Including ambiguous cases enhances model robustness.

Disagreement does not necessarily indicate poor annotation quality.

Abstract

Since state-of-the-art approaches to offensive language detection rely on supervised learning, it is crucial to quickly adapt them to the continuously evolving scenario of social media. While several approaches have been proposed to tackle the problem from an algorithmic perspective, so to reduce the need for annotated data, less attention has been paid to the quality of these data. Following a trend that has emerged recently, we focus on the level of agreement among annotators while selecting data to create offensive language datasets, a task involving a high level of subjectivity. Our study comprises the creation of three novel datasets of English tweets covering different topics and having five crowd-sourced judgments each. We also present an extensive set of experiments showing that selecting training and test data according to different levels of annotators' agreement has a strong effect on classifiers performance and robustness. Our findings are further validated in cross-domain experiments and studied using a popular benchmark dataset. We show that such hard cases, where low agreement is present, are not necessarily due to poor-quality annotation and we advocate for a higher presence of ambiguous cases in future datasets, particularly in test sets, to better account for the different points of view expressed online.