AI Memory Boundary Fit: De-identified Study Data

De-identified participant-level data for the manuscript Remembering or monitoring? Contextual fit of simulated AI memory cues and consumer trust (Huang, Huang, & Guo; prepared for submission to Frontiers in Psychology).

Six interactive scenario experiments (total N = 2,140) test how an AI assistant's memory type (no memory / relevant memory / intrusive memory) shapes consumer trust, purchase intention, data-sharing willingness, and loyalty, and whether the effects operate through Memory Boundary Fit (MBF) — the degree to which remembered information fits the consumer's contextual boundary for the current interaction.

Datasets

File	N	Design	Key measures
data/study1_memory_data.csv	264	Memory type (3 cells)	Trust, purchase intention, continued use, MBF
data/study2_memory_data.csv	289	Memory type (3 cells)	Trust, MBF, relational warmth, surveillance concern
data/study3_memory_data.csv	410	Memory type × product sensitivity (6 cells)	Trust
data/study4_memory_data.csv	516	Memory type × control interface (6 cells)	Trust, MBF
data/study5_memory_data.csv	318	Memory type (3 cells)	Transactional, data-sharing, and relational engagement; MBF
data/study6_memory_data.csv	343	Memory type (3 cells)	Privacy trust, loyalty, opt-in intention, platform trust, MBF
validation/mbf_validation_data.csv	234	Item-level validation sample	MBF items, IUIPC privacy concern, relational warmth, surveillance concern, trust

Additional files:

• data/demographics_by_participant.csv: participant-level demographics for the pooled sample (age, gender, student status, testing channel only)
• data/demographics_summary.csv: summary demographics by study
• data/data_dictionary.md: variable descriptions for every dataset

Notes on the data

• All multi-item constructs were measured on seven-point scales; item-level responses are provided where available, and composite scores are the mean of their items.
• Data were collected between March and June 2026 under a protocol approved by the Institutional Review Board of Zhejiang Business College (Protocol No. ZJBCSCCT-2026-EXP007). Participants provided electronic informed consent and were debriefed.
• The datasets are fully de-identified: the only person-level fields are age, gender, student status, and testing channel. Raw screening logs and conversation logs are withheld under the IRB's privacy conditions.

Availability of other materials

The stimulus protocol and supplementary tables accompany the article as Supplementary Material. Analysis code and exact statistics are available from the corresponding author upon reasonable request.

License and citation

The datasets are released under CC BY 4.0 (see LICENSE). If you use these data, please cite the associated article (see CITATION.cff).