Are We Studying the Same Filter Bubble? A Conceptual Framework of the Empirical Literature

This repository contains the supplementary material accompanying the paper “Are We Studying the Same Filter Bubble? A Conceptual Framework of the Empirical Literature”, published in the Joint Proceedings of the ACM UMAP Workshops 2026, UMAP 2026, June 8–11, 2026, Gothenburg, Sweden. 

Abstract

Since its introduction, the “filter bubble” concept has sparked extensive multidisciplinary research, yet empirical studies differ substantially in how they define, observe, and operationalize the phenomenon. This conceptual ambiguity makes it difficult to compare findings or synthesize evidence across studies. In this paper, we examine how filter bubbles have been operationalized in the empirical literature. Drawing on a grounded theory analysis of 40 empirical studies, we inductively derive a conceptual framework that distinguishes the components involved in filter bubble formation, the information exchanges between them, the components where filter bubbles are observed, and the components treated as shaping or mitigating them. Applying this framework, we identify three recurring archetypes of filter bubble operationalization: the algorithmic exposure bubble, the user-behavior-narrowing bubble, and the preference-reinforcement bubble. These archetypes show that empirical studies often examine different phenomena under the same label. The framework provides a structured vocabulary for categorizing existing studies, clarifying the relationship between filter bubbles and adjacent concepts such as echo chambers, and identifying theoretical and methodological blind spots in the current literature.

Supplementary material

The supplementary material consists of the following files:

CodedSet.pdf: List of the 40 papers included in the grounded theory analysis.

CodedSetStats.pdf: Descriptive statistics of the 40 papers included in the grounded theory analysis.

DataCollectionCodingSet.pdf: Documentation of the procedure used to search and select the set of papers from which the 40 coded papers were selected.

DataCollectionValidationSet.pdf:Documentation of the procedure used to search and select the set of papers from which the 7 validation papers were selected.

FrameworkAppliedToCodedSet.pdf: Visual application of the conceptual framework to the 40 coded papers, including the corresponding filter bubble archetypes.

FrameworkAppliedToValidationSet.pdf: Visual application of the conceptual framework to the 7 validation papers, including the corresponding filter bubble archetypes.

ValidationSet.pdf: List of the 7 papers included in the validation set.

These materials provide additional transparency on the data collection, theoretical sampling, coding, framework development, and validation procedures reported in the paper.