Evaluating the Ability of Computationally Extracted Narrative Maps to Encode Media Framing

Sebastián Concha Macías,Brian Keith Norambuena

2024-05-04

Abstract:Narratives serve as fundamental frameworks in our understanding of the world and play a crucial role in collaborative sensemaking, providing a versatile foundation for sensemaking. Framing is a subtle yet potent mechanism that influences public perception through specific word choices, shaping interpretations of reported news events. Despite the recognized importance of narratives and framing, a significant gap exists in the literature with regard to the explicit consideration of framing within the context of computational extraction and representation. This article explores the capabilities of a specific narrative extraction and representation approach -- narrative maps -- to capture framing information from news data. The research addresses two key questions: (1) Does the narrative extraction method capture the framing distribution of the data set? (2) Does it produce a representation with consistent framing? Our results indicate that while the algorithm captures framing distributions, achieving consistent framing across various starting and ending events poses challenges. Our results highlight the potential of narrative maps to provide users with insights into the intricate framing dynamics within news narratives. However, we note that directly leveraging framing information in the computational narrative extraction process remains an open challenge.

Computation and Language,Information Retrieval

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve

This paper explores the capability of computationally extracted narrative maps in encoding media framing. Specifically, the study aims to answer the following two key questions:

1. **RQ1: Can the narrative map extraction algorithm accurately capture the distribution of frames in the dataset?**
- This question focuses on whether the extraction algorithm can accurately reflect the proportion and distribution of different frames in the news dataset.

2. **RQ2: Can the narrative map extraction algorithm generate representations with consistent framing?**
- This question focuses on whether the extraction algorithm can maintain framing consistency between different starting and ending events when generating narrative maps.

### Research Background and Motivation

- **Importance of Narratives**: Narratives are fundamental frameworks for understanding the world and are crucial for collaborative sensemaking.
- **Role of Media Framing**: Media framing influences public perception and opinion through specific word choices and plays a significant role in the interpretation of news reports.
- **Research Gap**: Although the importance of narratives and framing is widely recognized, there is relatively little research that explicitly considers framing elements in the computational extraction and representation of narratives.

### Methods and Data

- **Dataset**: The study uses the Gun Violence Frame Corpus provided by Liu et al., which contains 1300 English news articles from various U.S. news sources reporting on gun violence issues in 2018.
- **Method**: The narrative extraction method developed by Keith and Mitra is used to generate narrative maps. This method generates a directed acyclic graph (DAG) through an optimization process, where nodes represent events and edges represent causal relationships and temporal sequences between events.

### Main Findings

- **Capturing Frame Distribution**: The results indicate that the extraction algorithm can effectively capture the frame distribution of the dataset, with an average Jensen-Shannon divergence of only 0.0218, indicating a high consistency between the generated narrative maps and the dataset's frame distribution.
- **Frame Consistency Issues**: Although the algorithm can capture the overall frame distribution, the frame consistency between different events in the generated narrative maps is poor. For example, some maps show significant inconsistencies when dealing with different frames such as school safety, political debate, and cultural impact.

### Significance and Impact

- **Research Significance**: This study highlights the importance of considering framing in computational narrative extraction and points out the current technology's limitations in maintaining frame consistency.
- **Practical Applications**: For practitioners such as journalists, policymakers, and analysts, computational narrative extraction technology has potential value in quickly identifying dominant frames and assessing their relative importance. However, attention should be paid to the issue of frame inconsistency in individual narrative maps.

### Conclusion

- **Main Conclusion**: The narrative map extraction algorithm performs well in capturing the frame distribution of the dataset but faces challenges in generating narrative maps with consistent framing.
- **Future Directions**: Future research should explore new methods and techniques to more effectively maintain frame consistency during the extraction process, thereby improving the accuracy and contextual understanding of the extracted data.