Study 1: Single Article Satirical Preference Flow

This repository hosts the first prototype used in our experimental research on reader preferences for satirical news content.

Description

In this study, participants read only one version of each article, randomly assigned to one of three satirical prompt styles. Each participant is also assigned to either a transparent or non-transparent condition releated to the disclosure of AI-usage.

Each article is followed by Likert-scale questions to evaluate things like humor, engagement, comprehension, and expectations.

Contents

• app/ — Flask-based app code
• templates/ — HTML for article pages and questionnaires
• static/ — Frontend assets
• manipulated_articles.csv — Dataset of articles split across prompt conditions

Requirements

Note: To run all the scripts in the repository one would need access to both a NewsCatcher API key and a OpenAI API key. This is only needed if you want to test with fresh rewritten articles.

Releated to paper: Using Large Language Models to 'Lighten the Mood': Satirically Reframing News Recommendations to Reduce News Avoidance.

Running the App

pip install -r requirements.txt
python app.py