Questionnaire Dataset on the Effects of Student–AI Trust on Learning Engagement and Learning Satisfaction in AI-Based Educational Technology

Several items were subsequently revised to better fit AI-supported learning contexts.The dataset is provided as one Microsoft Excel file named data.xlsx, with a file size of approximately 184 KB. The worksheet contains 1,106 data rows and 40 columns. Each row represents one valid questionnaire response from an individual participant.

The first column, labelled “序号,” is a respondent sequence identifier used solely for data organization and does not contain personally identifiable information. The remaining 39 columns contain responses to questionnaire items. Specifically, CT1–CT5 represent cognitive trust, AT1–AT5 represent affective trust, FT1–FT4 represent functional trust, ET1–ET5 represent environmental trust, CE1–CE5 represent cognitive engagement, AE1–AE5 represent affective engagement, BE1–BE5 represent behavioral engagement, and LS1–LS5 represent learning satisfaction.

Student–AI trust therefore comprises 19 items across four dimensions, learning engagement comprises 15 items across three dimensions, and learning satisfaction comprises five items.All questionnaire responses are recorded as integers on five-point Likert scales. For the student–AI trust items, values range from 1 (“strongly disagree”) to 5 (“strongly agree”). For the learning engagement and learning satisfaction items, values range from 1 (“very inconsistent”) to 5 (“very consistent”).

Higher values indicate higher levels of the corresponding trust, engagement, or satisfaction construct. The values are ordinal response codes rather than physical measurement units. The dataset contains no blank or missing responses in the 39 questionnaire-item columns.

No imputed values, estimated values, or replacement codes for missing data are included. All item responses fall within the expected range of 1 to 5, and no out-of-range item values were identified.The uploaded file contains the valid questionnaire responses used in the study. The data were organized and coded in Microsoft Excel and subsequently analysed using exploratory factor analysis, confirmatory factor analysis, structural equation modelling, and bias-corrected mediation analysis in Mplus.

The dataset can be reused for descriptive analysis, reliability testing, construct validity assessment, correlation analysis, factor analysis, structural equation modelling, mediation analysis, and methodological replication. Microsoft Excel, LibreOffice Calc, R, Python, SPSS, Stata, or similar statistical software can be used to open or analyse the file. The dataset does not include participants’ names, student numbers, contact information, or other direct personal identifiers.