Longitudinal Secondary Dataset: Generative AI Governance, Enterprise Adoption, and Curricular Outcomes in AACSB/EQUIS Business Schools (2021–2025)
Description
This dataset provides an audited, balanced 4-year longitudinal secondary panel of 30 flagship AACSB- and EQUIS-accredited public business faculties across the United States (n = 25) and the United Kingdom (n = 5), observed across four consecutive academic cycles: 2021–2022 (pre-diffusion baseline), 2022–2023 (exogenous technology shock), 2023–2024 (formal policy codification), and 2024–2025 (institutional maturation and enterprise scale), yielding 120 balanced institutional-year observations. The primary objective of this repository is to support empirical investigations into higher education governance, curricular adaptation, and instructional outcomes following the widespread emergence of generative artificial intelligence (GenAI) in management education, specifically accompanying research submitted to The International Journal of Management Education (IJME). In strict adherence to open-science integrity and empirical transparency, this dataset contains zero simulated, synthetic, or calibrated data points. All records are compiled exclusively from verifiable administrative archives and documentary sources: 1. Institutional AI Oversight Rigor Index (oversight_score, 0–4 scale): Hand-coded via systematic documentary audit of publicly available academic senate resolutions, honor council codes, and business faculty syllabus directives across four dimensions: mandatory syllabus usage disclosure, process-based oral defense reservations, formal academic misconduct referral pathways, and dedicated ethical AI training modules. 2. Pedagogical Stance & Enterprise Infrastructure: Coded classifications of primary syllabus posture (prohibitive, permissive with attribution, integrated) and official deployment of enterprise-grade AI sandboxes (ai_enterprise_sandbox, 0/1) verified through institutional CIO/IT disclosures. 3. Objective Educational Outcome Proxy: Degree completion shares in Management Science and Quantitative Methods (CIP Code 52.13) relative to total business degree completions (CIP Code 52), extracted directly from the National Center for Education Statistics (NCES) IPEDS Completions Survey and the Higher Education Statistics Agency (HESA/Jisc). 4. Administrative Baseline Controls: Full-Time Equivalent Enrollment (enroll_fte), Student-to-Faculty Ratios (sf_ratio), and Instructional Expenditures per FTE (instructional_exp_per_fte_usd) extracted from IPEDS Fall Enrollment, Institutional Characteristics, and Finance surveys, and corresponding HESA statistical registries. The workbook is structured into four functional worksheets: (1) 'ReadMe_and_Metadata' detailing deposit protocols and econometric replication syntax (Stata, R, Python); (2) 'Longitudinal_Panel_Data' containing the complete sectioned panel matrix; (3) 'Policy_Audit_and_Citations' providing live public URLs and verbatim excerpts justifying every hand-coded score; and (4) 'Codebook_and_Rubric' defining all variable operationalizations and scoring criteria.
Files
Steps to reproduce
1. ARCHIVAL EXTRACTION & DATASET ARCHITECTURE The primary empirical dataset is in 'Mendeley_Data_Institutional_AI_Governance_2021_2025.xlsx'. It comprises a balanced 4-year longitudinal panel of 30 AACSB/EQUIS public business schools (25 US, 5 UK) observed across four cycles: 2021-2022 (pre-diffusion baseline), 2022-2023 (technology shock), 2023-2024 (formal codification), and 2024-2025 (enterprise adoption), totaling 120 institutional-year observations. 2. VERIFICATION OF ADMINISTRATIVE COVARIATES - US Institutions: Cross-check IDs (IPEDS UNITID) via the NCES IPEDS Data Center. Covariates—Full-Time Equivalent Enrollment (enroll_fte), Student-Faculty Ratio (sf_ratio), and Instructional Expenditure per FTE (instructional_exp_per_fte_usd)—derive from IPEDS Fall Enrollment, Institutional Characteristics, and Finance surveys. - UK Institutions: Verify figures against the Higher Education Statistics Agency (HESA/Jisc) using official UKPRN codes. - Learning Outcome Proxy (quant_degree_share_pct): Replicate via the IPEDS Completions Survey. Calculate the ratio of conferred degrees under CIP Code 52.13 (Management Science & Quantitative Methods) to total business degrees conferred (CIP Code 52), multiplied by 100. 3. AUDIT OF INSTITUTIONAL AI GOVERNANCE SCORES Open sheet 'Policy_Audit_and_Citations' to verify AI oversight rigor scores (oversight_score, 0-4): - Navigate to the provided public URLs for each university teaching center or academic senate archive. - Verify verbatim excerpts against the 4-point rubric in sheet 'Codebook_and_Rubric': * Criterion 1 (C1): Mandatory syllabus disclosure of GenAI tools and prompt queries (+1). * Criterion 2 (C2): Reservation of faculty rights to require oral defenses or process tracking (+1). * Criterion 3 (C3): Formal referral pathway for unauthorized AI generation to honor boards (+1). * Criterion 4 (C4): Mandatory student ethics/AI integrity training modules (+1). - Confirm enterprise deployment (ai_enterprise_sandbox = 1) via official CIO/IT disclosures. 4. STATISTICAL MODEL ESTIMATION To replicate the Two-Way Fixed Effects (TWFE) panel regressions: - In Stata: import excel "Mendeley_Data_Institutional_AI_Governance_2021_2025.xlsx", sheet("Longitudinal_Panel_Data") firstrow clear encode academic_year, gen(year_id) xtset inst_id year_id xtreg quant_degree_share_pct oversight_score ai_enterprise_sandbox log(enroll_fte) sf_ratio log(instructional_exp_per_fte_usd) i.year_id, fe vce(cluster inst_id) - In R: library(readxl); library(plm) df <- read_excel("Mendeley_Data_Institutional_AI_Governance_2021_2025.xlsx", sheet = "Longitudinal_Panel_Data") pdata <- pdata.frame(df, index = c("inst_id", "academic_year")) model <- plm(quant_degree_share_pct ~ oversight_score + ai_enterprise_sandbox + log(enroll_fte) + sf_ratio + log(instructional_exp_per_fte_usd), data = pdata, model = "within", effect = "twoways") summary(model)