From "Black Box" to Trusted Partner: The Role of Explainability and User Control in Digita

132 postpartum women participated in an 8-week longitudinal experiment between early 2026 and May 2026. The data captures the effects of AI system Explainability and User Control on trust, autonomy, and exercise adherence for postpartum depression. Feng, Aojie submitted this anonymized dataset to Harvard Dataverse, providing quantitative evidence from a study approved by Soochow University in mainland China.

Use Cases

• Modeling the causal impact of AI interface design on clinical outcomes based on the 2x2 factorial design.
• Analyzing longitudinal trends in depression symptoms based on Edinburgh Postpartum Depression Scale scores collected at four intervals.
• Investigating the relationship between psychological constructs like Perceived Autonomy and objective Behavioral Performance metrics like Adherence Rate Percent.
• Benchmarking digital health intervention efficacy within an Eastern cultural and healthcare context based on the study's geographic scope.

Strengths

• Data is longitudinal, tracking 132 participants over a continuous 8-week intervention period.
• Study design employs a rigorous 2x2 between-subjects factorial design with four randomized experimental groups.
• Dataset integrates multi-dimensional attributes including demographic profiles, clinical outcomes, HCI metrics, and behavioral performance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-05-29 11:43:14; freshness should be verified as the data pertains to a future period.

Provenance

Source

Harvard Dataverse, submitted by Feng, Aojie.

Collection Method

Empirical data from a randomized controlled trial with a 2x2 between-subjects factorial design.

Time Range

Data collected between early 2026 and May 2026.

Freshness

Last updated 2026-05-29 11:43:14.

Geography

Study conducted in mainland China.