A Cross-sectional Dataset on Technology Acceptance, Academic AI-Integration, Behavioral and Self Directed Learning Readiness among Law Students in Bangladesh

Abstract

This dataset contains cross-sectional survey data from 412 enrolled law students at a private university in Bangladesh, collected between 10 January and 10 July 2026 using a structured online questionnaire. Participants were recruited through purposive sampling. The dataset was developed to support research on generative-AI adoption in legal education by bringing together measures of technology acceptance, academic AI use, and self-directed learning readiness within the same respondent population. The questionnaire included demographic characteristics and AI-tool usage; the Technology Acceptance Model (TAM), comprising Perceived Usefulness and Perceived Ease of Use; four AI Literacy & Ethics items addressing formal AI training, institutional AI-policy awareness, data-privacy concerns, and concern regarding bias in AI-generated content; the 24-item Academic AI Usage Scale (AAIUS); and the 40-item Self-Directed Learning Readiness Scale (SDLRS), comprising Self-Management, Self-Control, and Desire for Learning. TAM and AAIUS were administered using an adapted three-point Disagree–Neutral–Agree response format, coded 0–2. SDLRS retained its five-point Strongly Disagree–Strongly Agree response format, coded 1–5; three originally reverse-worded items were rephrased before administration to align their direction with the remaining scale items and therefore did not require reverse scoring. The release includes de-identified source-level, cleaned, and analysis-ready respondent-level datasets; a detailed variable codebook and scoring reference; response-frequency tables; missing-data documentation; item- and composite-level descriptive statistics; scale-level internal-consistency estimates; AAIUS descriptive-band and SDLRS readiness-classification distributions; and a composite-level Spearman correlation matrix. A reproducible data-processing script is also provided. Direct identifiers and original submission timestamps were removed before synthetic respondent identifiers (NewID) were assigned. Two optional open-ended items were collected, but their verbatim responses are withheld from the public release because of potential disclosure risk; only response-completion counts are provided. The dataset may support descriptive, correlational, subgroup, multivariable, and subsequent psychometric research on generative-AI use in higher education, particularly in legal education, and may facilitate replication and comparison across institutions, disciplines, and jurisdictions.