Best-Worst Scaling More Reliable than Rating Scales: A Case Study on Sentiment Intensity Annotation

TL;DR

This study demonstrates that Best-Worst Scaling (BWS) provides more reliable sentiment intensity annotations than traditional rating scales, with a systematic comparison confirming BWS's superior consistency.

Contribution

The paper presents the first systematic comparison between BWS and rating scales, establishing BWS as a more reliable annotation method for sentiment intensity.

Findings

BWS yields significantly more reliable annotations than rating scales.

Both methods used the same total number of annotations.

BWS improves inter- and intra-annotator consistency.

Abstract

Rating scales are a widely used method for data annotation; however, they present several challenges, such as difficulty in maintaining inter- and intra-annotator consistency. Best-worst scaling (BWS) is an alternative method of annotation that is claimed to produce high-quality annotations while keeping the required number of annotations similar to that of rating scales. However, the veracity of this claim has never been systematically established. Here for the first time, we set up an experiment that directly compares the rating scale method with BWS. We show that with the same total number of annotations, BWS produces significantly more reliable results than the rating scale.