SERP data from controversial queries on Google and Bing

Data for the forthcoming publication 'Contested Components: Studying Interface Enrichment as a Form of Content Moderation on Google and Bing'.

Datasets contain information on SERP components for Google Bing when querying 2000 controversial and 914 non-controversial questions.

Files include:

question_data.csv: Information on questions sourced from 4chan and leftychan boards in November 2024. Columns include the counts per board (/fit/, /b/, /pol/. /int/, /k/, /lgbt/, and /leftypol/), categorization as controversial/non-controverial, and toxicity scores determined by Perspective API.

serp_components.csv: Information on the SERP data gathered using Zoekplaatje. Collected on 24 November 2024.

screenshots.zip: Screenshots of all SERPs. Note that at times, expanding the AI Overview box on Google resulted in the search bar overlaying the generated text.

component_analysis.ipynb: Code for analyzing the data.