From Ground Truth to Measurement: A Statistical Framework for Human Labeling
2026-04-08 • Artificial Intelligence
Artificial IntelligenceComputation and LanguageMachine Learning
AI summaryⓘ
The authors explain that when humans label data for machine learning, the labels can vary due to confusing examples or different opinions, not just mistakes. They created a method to break down these variations into understandable parts like how hard an item is, individual biases, random mistakes, and how well people agree. Their approach helps distinguish whether there's one true answer or multiple valid ones. They tested this on language data and found evidence for all these factors, suggesting their method can improve how we understand and work with labeled data.
Supervised machine learningHuman annotationLabel noiseMeasurement errorAnnotator biasInstance difficultyNatural language inferenceRelational alignmentData-centric machine learningStatistical modeling
Authors
Robert Chew, Stephanie Eckman, Christoph Kern, Frauke Kreuter
Abstract
Supervised machine learning assumes that labeled data provide accurate measurements of the concepts models are meant to learn. Yet in practice, human labeling introduces systematic variation arising from ambiguous items, divergent interpretations, and simple mistakes. Machine learning research commonly treats all disagreement as noise, which obscures these distinctions and limits our understanding of what models actually learn. This paper reframes annotation as a measurement process and introduces a statistical framework for decomposing labeling outcomes into interpretable sources of variation: instance difficulty, annotator bias, situational noise, and relational alignment. The framework extends classical measurement-error models to accommodate both shared and individualized notions of truth, reflecting traditional and human label variation interpretations of error, and provides a diagnostic for assessing which regime better characterizes a given task. Applying the proposed model to a multi-annotator natural language inference dataset, we find empirical evidence for all four theorized components and demonstrate the effectiveness of our approach. We conclude with implications for data-centric machine learning and outline how this approach can guide the development of a more systematic science of labeling.