Campaigning through the lens of Google: A large-scale algorithm audit of Google searches in the run-up to the Swiss Federal Elections 2023

2025-07-08

AI Summary

This study conducts the first large-scale audit of Google’s search result presentation for Swiss federal election candidates in 2023, examining algorithmic bias along gender and party lines in both textual and image results. Method: Leveraging multi-round search queries, the authors employ automated gender/party classification, sentiment analysis, and pre-post comparison to assess representational disparities. Contribution/Results: Textual results exhibit significant male bias; image results reinforce stereotypical “joyful” portrayals of right-wing female candidates. Critically, presentation patterns—including ranking position, sentiment valence, and visual framing—robustly predict candidates’ actual vote shares. The study thus uncovers latent algorithmic bias mechanisms in electoral information ecosystems and pioneers a novel linkage between algorithmic auditing and electoral outcome prediction. By empirically demonstrating how search interfaces shape political visibility and influence electoral performance, it provides foundational evidence for platform accountability and evidence-based digital democratic governance.

Technology Category

• Search and Optimization: Algorithm Configuration
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Computer Vision: Bias, Fairness & Privacy

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Search engines like Google have become major sources of information for voters during election campaigns. To assess potential biases across candidates' gender and partisan identities in the algorithmic curation of candidate information, we conducted a large-scale algorithm audit analyzing Google's selection and ranking of information about candidates for the 2023 Swiss Federal Elections, three and one week before the election day. Results indicate that text searches prioritize media sources in search output but less so for women politicians. Image searches revealed a tendency to reinforce stereotypes about women candidates, marked by a disproportionate focus on stereotypically pleasant emotions for women, particularly among right-leaning candidates. Crucially, we find that patterns of candidates' representation in Google text and image searches are predictive of their electoral performance.

Problem

• Assess gender and partisan biases in Google search results
• Analyze algorithmic curation of Swiss election candidate information
• Examine search result impact on electoral performance

Innovation

• Large-scale algorithm audit of Google searches
• Analyzed gender and partisan biases in search results
• Linked search representation to electoral performance