Anonymised Round-Level Dataset: Delphi Study on AI Integration in Kazakhstan's E-Government Project Management

This dataset contains the anonymised, round-level (aggregate) data underlying Tables 1–6 of the manuscript "A Delphi–Monte Carlo Framework for Probabilistic KPI Estimation in Public Sector AI Projects: Evidence from Kazakhstan's E-Government" (Ismailova et al., 2026). The dataset was generated through a three-round modified Delphi study with N = 20 domain experts from Kazakhstan's digital government and public-sector project management ecosystem (5–14 May 2026). Included files: - table1_panel_demographics.csv — aggregate panel composition - table2_r2_toe_importance.csv — Round 2 TOE factor importance ratings - table3_band_frequencies.csv — Round 2 categorical band-frequency distributions and band-to-midpoint mapping - midpoint_vectors.json — the eight 20-element numeric midpoint vectors used as direct input to the Monte Carlo simulation (reproduces Table 6) - table4_r3_kpi_agreement.csv — Round 3 KPI agreement statistics - table5_r3_toe_agreement.csv — Round 3 TOE factor agreement ratings - qualitative_themes_r3.csv — frequency counts of Round 3 open-ended themes - Appendix tables A1–A3 (.docx) Individual-level expert responses are withheld to protect participant confidentiality, consistent with the Data Availability Statement of the associated manuscript. For the simulation code that reproduces Table 6 and Figures 1–2 from this dataset, see the companion record: https://doi.org/10.5281/zenodo.20687853 (or same record if combined).