Reliability in content analysis: The case of semantic feature norms classification

Area of Science:

• Cognitive Psychology
• Computational Linguistics
• Psycholinguistics

Background:

• Semantic feature norms are crucial in cognitive psychology for understanding concept representation.
• Current methods for annotating semantic features lack consistency, posing methodological challenges.
• Inconsistent content analysis of semantic features may compromise theoretical conclusions.

Purpose of the Study:

• To review existing semantic feature norm datasets and taxonomies.
• To provide theoretical and methodological insights into content analysis for semantic features.
• To propose a standardized methodology for reliable and replicable semantic feature annotation.

Main Methods:

• Review of annotated semantic feature norm datasets and content analysis taxonomies.
• Theoretical and methodological analysis of content analysis procedures.
• Application of content analysis to a new semantic feature dataset, focusing on taxonomy structure, category description, coder training, and coding scheme sustainability.

Main Results:

• Methodological guidelines for semantic feature classification are presented.
• A revised and adaptable taxonomy for classifying semantic features of both concrete and abstract concepts is proposed.
• A new dataset of annotated semantic feature norms is provided.

Conclusions:

• Standardized content analysis methods are essential for reliable semantic feature norm annotation.
• The developed guidelines and taxonomy enhance the quality and replicability of semantic feature research.
• The new annotated dataset facilitates future investigations into semantic cognition.