Poster #118 - When Transparency Works: Disclosure Timing and Public Responses to Algorithmic Decisions

• In Event: Saturday Poster Luncheon

Abstract

Governments are rapidly automating high-stakes administrative decisions in welfare, healthcare, and tax systems. To address public concerns, jurisdictions worldwide — including the EU, the United States, and China — have adopted governance frameworks centered on algorithmic transparency and human review of appeals. Yet evidence on these safeguards remains mixed: some studies find that transparency increases trust, while others show it generates resistance and reduces program effectiveness. Why does the same instrument sometimes function as a remedy and sometimes as a liability? We examine one overlooked explanation: disclosure timing. We test whether learning about algorithmic involvement before versus after a decision helps explain these divergent findings and conditions the effectiveness of transparency and human review.
The study also addresses a broader question in policy design. Research on policy instruments has examined layering, sequencing, and combinations across policy episodes, but less attention has been paid to whether temporal parameters within a single policy episode shape instrument effectiveness. Procedural justice theory suggests they may: information available before an outcome is known shapes expectations and interpretations of the process. Pre-decision disclosure may establish a frame in which transparency and review are understood as genuine safeguards. Post-decision disclosure, by contrast, may cause the same measures to be perceived as retrospective justification — formally present but procedurally too late to be meaningful.
We test these expectations using a pre-registered 2×2×2 between-subjects experiment in the context of automated social assistance eligibility review. Participants are randomly assigned to conditions varying disclosure timing (before vs. after the decision), algorithmic transparency (detailed vs. minimal explanation), and human review (staff-handled vs. automated appeal). The primary outcome is procedural justice, measured through two validated dimensions: decision quality (perceived accuracy and fairness) and treatment quality (perceived respect and information provision). All scales demonstrate strong reliability and confirmed discriminant validity.
Three findings emerge. First, pre-decision notification significantly improves evaluations of procedural justice, institutional legitimacy, and algorithm acceptance. Among the design features tested, notification timing appears both highly consequential and administratively feasible. Second, disclosure timing conditions the returns to transparency. Detailed explanations substantially improve citizen evaluations when individuals are informed before the decision, but these gains weaken when disclosure occurs only afterward. Transparency is not self-executing; its effectiveness depends on when citizens first learn of automated involvement. Third, human review of appeals does not significantly improve procedural justice under either timing condition, though it modestly increases institutional trust — suggesting it functions more as an accountability signal than as a process-level safeguard.
These findings identify disclosure timing as a reconciling variable for contradictory evidence on algorithmic transparency and open a research agenda on how the temporal structure of policy instruments shapes their joint effectiveness. For practitioners, the results point to a low-cost reform: requiring pre-decision notification may substantially enhance the returns on existing transparency investments without new technology or significant administrative burden.

Author

• CHEN SIQI, University of Chinese Academy of Sciences
  Presenting Author