Moral rhetoric in discrete choice models: a Natural Language Processing approach

Area of Science:

• Social Psychology
• Computational Social Science
• Political Science

Background:

• Understanding moral motivations behind human actions is complex.
• Existing methods often analyze actions or stated preferences separately.
• Integrating verbal expressions with behavioral data offers a richer perspective.

Purpose of the Study:

• To develop and validate a new methodology for inferring moral motivations.
• To combine choice data with text data for enhanced behavioral analysis.
• To explore the role of moral rhetoric in political decision-making.

Main Methods:

• Utilizing Natural Language Processing (NLP) to extract moral values (moral rhetoric) from text.
• Employing Moral Foundations Theory as the framework for moral value identification.
• Integrating extracted moral rhetoric as input into Discrete Choice Models.
• Case study analysis of voting and party defection in the European Parliament.

Main Results:

• Moral rhetoric demonstrates significant explanatory power in modeling voting behavior.
• The proposed method successfully integrates textual moral values with choice data.
• Insights into moral behavior are gained by analyzing both words and actions.

Conclusions:

• The novel method provides a robust approach to inferring moral motivations.
• Moral rhetoric is a valuable predictor of political behavior.
• Future research can extend this methodology to other domains of human action.