Data and reproducibility code for "Interdisciplinary teaching experience and satisfaction in postgraduate design education: an explanatory and predictive analysis"

Description This repository contains the de-identified analysis dataset, variable codebook, and reproducibility code supporting the study “Interdisciplinary teaching experience and satisfaction in postgraduate design education: an explanatory and predictive analysis.” The study analysed survey responses from 154 postgraduate design students who participated in a multi-course interdisciplinary teaching reform. The quantitative instrument included five teaching-experience facets: Content Quality, Supportive Learning Climate, Perceived Utility, Learning Engagement, and Instructional Credibility, together with one global item assessing Overall Teaching Satisfaction. The repository contains the participant-level variables required to reproduce the quantitative analyses reported in the manuscript. To protect participant confidentiality, platform-generated metadata and potentially identifying information have been excluded. The public dataset does not contain IP addresses, survey-source metadata, precise submission timestamps, response-duration metadata, original platform sequence numbers, names, student identification numbers, or open-ended responses to Q25. The accompanying code reproduces the principal descriptive, reliability, empirical-distinctiveness, explanatory, robustness, and predictive analyses reported in the revised manuscript. These include Cronbach’s alpha, HTMT, variance inflation factors, exploratory dimensionality diagnostics and parallel analysis, ordinary least-squares models, BCa bootstrap confidence intervals, HC3 robust inference, Breusch–Pagan diagnostics, alternative integrated-score sensitivity analyses, proportional-odds diagnostics, multinomial-logit sensitivity analysis, and repeated nested cross-validation of the artificial neural network. Predictive validation uses 20 repetitions of 10-fold outer cross-validation. ANN hyperparameters are selected exclusively within five-fold inner cross-validation. The principal random seed used for reproducibility is 20260810. The codebook provides variable definitions, response coding, translated questionnaire wording, facet membership, scoring procedures, and the theoretical basis and contextual adaptation of questionnaire items. This repository is intended to provide the minimum de-identified dataset and computational materials required to reproduce and scrutinise the quantitative findings of the associated manuscript.