Dark Patterns in Indonesian Fintech Applications: Reconciled Audit Dataset and Recognition Survey Data (Rupiah***, Easy***, Ada***)

Description

Reconciled audit dataset (Round 1 + Round 2 intra-rater reliability re-coding, 14-day blind washout, 1-15 Aug 2026) and recognition survey data for the study 'Dark Patterns in Indonesian Fintech Applications: An Audit and a Study of User Recognition.' Contains 54 coded UI-element records from a structured audit of three anonymized Indonesian fintech applications (Rupiah***, Easy***, Ada***), spanning paylater/BNPL, e-wallet, and OJK-licensed peer-to-peer lending, each coded against a ten-category dark-pattern taxonomy (Nagging, Obstruction, Sneaking/Hidden Costs, Interface Interference, Forced Action, Urgency & Scarcity, Social Proof, Confirmshaming, Trick Questions, Privacy Zuckering). Round 1 (1 Aug 2026) and Round 2 (15 Aug 2026, blind re-coding) intra-rater reliability was Krippendorff's alpha = 0.522 (MASI distance); the 20 items that disagreed were resolved through a documented reconciliation pass (six downgraded to 'no dark pattern'). Also includes Phase 2 recognition survey data: n = 124 valid responses to a screenshot-based, signal-detection-scored recognition instrument (8 scored dark-pattern stimuli + 3 neutral controls), fielded via Google Forms. The audit covers the registration-to-cancellation user flow under a fixed stopping rule; all optional permission requests were declined and the walkthrough terminated before any national-ID (KTP) upload.

Steps to reproduce

1) Open 3_Audit_Codebook.xlsx to review the ten-category coding scheme and evidence rules. 2) Open 1_Audit_Log_Final_Reconciled.xlsx for all 54 coded UI-element records, including both the original Round 1 label and the FINAL post-reconciliation label (with rationale) for each element. 3) Cross-reference the FINAL Category Label(s) column against the codebook to reproduce the prevalence matrix reported as Table 1 in the accompanying manuscript. 4) Open 2_Phase2_Survey_Data.xlsx for the n = 124 recognition-survey responses; see its Column Dictionary and Notes sheets for the stimulus-to-category mapping and the Sneaking-stimulus exclusion rationale. 5) See README.md for full version history and a file guide.

Institutions

• Binus University

  Jakarta, Jakarta

Categories

Licence