Multidimensional Algorithmic Trust in Financial Recommender Systems

This dataset contains the de-identified data underlying a between-subjects online experiment on multidimensional algorithmic trust in financial recommender systems. Study design. A 4 (recommendation source: default, personalized, social-proof, expert) × 2 (explanatory information: present, absent) between-subjects experiment. After viewing a financial product recommendation, participants reported their perceived accuracy, fairness, and explainability of the recommendation, together with satisfaction and continuance intention, each on multi-item 7-point scales. The data also include subjective financial literacy (self-rated items), objective financial literacy (multiple-choice knowledge items), experimental condition assignment, and demographic variables. Sample. 435 adults in South Korea with prior experience using digital financial services, recruited through a commercial online panel (Macromill Embrain) in January 2026. Files. Data are provided in CSV format (algorithmic_trust_FRS_experiment.csv). Ethics and access. The study was approved by the Seoul National University Institutional Review Board (Protocol No. 2511/003-003). Because participant consent and the approving ethical review did not provide for public data release, and because the data were collected under a commercial panel agreement precluding public redistribution, the files are shared under restricted access. De-identified data may be made available to qualified researchers for non-commercial academic purposes through Zenodo's access-request procedure, subject to these constraints and to approval by the corresponding author.