The Limits of Platform Observability: Statements of Reasons on Dating Apps under the Digital Services Act

Antoni Mut-Piña (Universitat de les Illes Balears), Rosa Barceló-Compte (Universitat de Barcelona), Carlotta Rigotti (Leiden University) & Eduard Fosch-Villaronga (Leiden University)

Download preprint (PDF) · September 2026

Research question: To what extent do the statements of reasons submitted by dating apps to the DSA Transparency Database under Article 17 DSA enable meaningful observability of their content moderation practices?

Abstract: When online platforms restrict user content or accounts, affected users and external observers often struggle to understand what was decided, why it was decided, and how moderation practices operate at scale. Article 17 of the Digital Services Act (DSA) seeks to address this opacity by requiring platforms to provide statements of reasons (SoRs) for content moderation decisions and to submit them to a public Transparency Database. This article examines whether these disclosures enable meaningful platform observability in the under-studied context of dating applications, where moderation governs access to intimate, relational, and identity-based forms of social participation. Drawing on the full population of 6,878,730 statements of reasons submitted by Badoo, Bumble, Hinge, Meetic, Tinder, Lovoo, and OkCupid over a six-month observation window, we estimate logistic regression models to identify when platforms provide substantive explanations. We find that 79.1% of moderation decisions include such an explanation, with explanation rates ranging from 7.1% to 98.9% across platforms. Rather than by a consistent norm of reason-giving, explanation provision is shaped by platform-level reporting infrastructures, automation pipelines, user-initiated reports, and account termination. Fully automated detection-and-decision pipelines are associated with higher explanation rates, while partial automation and user-initiated reports are associated with lower rates. These findings show that Article 17 DSA disclosures support sector-level observability, but only to a partial extent: they reveal patterns of standardization, platform-specific reporting practices, formalized transparency, and incomplete accountability, while leaving the underlying content, context, and potential discriminatory effects of moderation decisions largely outside view. The article contributes to Information Systems research by analyzing regulatory-imposed transparency databases as socio-technical infrastructures that shape what can, and cannot, be known about platform governance.

Data: The analysis draws on the DSA Transparency Database, which compiles metadata on content moderation decisions reported by online platforms operating in the EU under Article 24(5) DSA. Rather than drawing a sample, the study uses the full population of SoRs submitted by seven dating apps (Badoo, Bumble, Hinge, Meetic, Tinder, Lovoo, and OkCupid) between 1 December 2025 and 31 May 2026, yielding 6,878,730 row-level observations.

Method: Logistic regression models with platform fixed effects (reporting average marginal effects) examine whether explanation provision is associated with automated detection and decision-making, user-initiated notices, and the severity of the restriction. Two complementary dependent variables are used: Explanation Provided, whether the user-facing explanation names a specific content category or policy provision, and Vague Explanation, whether the platform reports the decision under a generic or residual category.