Quantifying and Attributing Polarization to Annotator Groups
This addresses the need for better inter-group analysis in annotation for subjective tasks, though it is incremental as it builds on existing work on annotation metrics.
The authors tackled the problem of measuring polarization among annotator groups in subjective tasks like hate speech detection, introducing a new metric and statistical test that works with imbalanced groups and multi-label settings, and applied it to four datasets to quantify polarization by race, religion, and education, finding strong racial polarization in hate speech tasks.
Current annotation agreement metrics are not well-suited for inter-group analysis, are sensitive to group size imbalances and restricted to single-annotation settings. These restrictions render them insufficient for many subjective tasks such as toxicity and hate-speech detection. For this reason, we introduce a quantifiable metric, paired with a statistical significance test, that attributes polarization to various annotator groups. Our metric enables direct comparisons between heavily imbalanced sociodemographic and ideological subgroups across different datasets and tasks, while also enabling analysis on multi-label settings. We apply this metric to three datasets on hate speech, and one on toxicity detection, discovering that: (1) Polarization is strongly and persistently attributed to annotator race, especially on the hate speech task. (2) Religious annotators do not fundamentally disagree with each other, but do with other annotators, a trend that is gradually diminished and then reversed with irreligious annotators. (3) Less educated annotators are more subjective, while educated ones tend to broadly agree more between themselves. Overall, our results reflect current findings around annotation patterns for various subgroups. Finally, we estimate the minimum number of annotators needed to obtain robust results, and provide an open-source Python library that implements our metric.