ANEST narrative-affect dataset (ANAD v1): A large-scale derived feature resource for quantifying narrative-affective...

Area of Science:

• Computational social science
• Affective science
• Narrative psychology

Background:

• Existing research often separates narrative structure and emotional valence.
• A comprehensive understanding of the interplay between life stories and feelings is lacking.

Purpose of the Study:

• To introduce a large-scale dataset and derived feature resource for analyzing narrative structure and affective valence in human-written narratives.
• To present the Narrative-Affect Discrepancy Index (NADI) as a computable indicator of the gap between narrative complexity and expressed affect.

Main Methods:

• Utilized a dataset of 351,734 English-language narratives from public online discussions about romantic relationships.
• Derived annotations included basic structural descriptors, a Level of Complexity (LoC) index, and a normalized affective polarity score.
• Developed the Narrative-Affect Discrepancy Index (NADI) to quantify the difference between narrative complexity and expressed affect on a 0-10 scale.

Main Results:

• The ANEST Narrative-Affect Dataset (ANAD v1) provides derived, non-identifiable feature representations.
• The dataset includes distributions of narrative complexity, affective polarity, and the NADI score.
• The NADI is presented as an operational indicator for research, not a validated psychological construct.

Conclusions:

• The ANAD v1 dataset and NADI offer a novel resource for computational social science, affective science, and narrative psychology.
• This resource facilitates empirical investigation into the complex relationship between narrative expression and emotional experience.
• The openly available dataset encourages further research into the structure and affect of personal storytelling.