IA-Recrut: data from a vignette experiment on AI-based application screening, anticipated discrimination and perceived fairness (Morocco, 2026)

This dataset contains the anonymised data from IA-Recrut, a vignette experiment conducted in Morocco (data collected from 4 January to 30 September 2026) on how job applicants perceive AI-based application screening. Participants were randomly assigned to experimental scenarios and reported their anticipated discrimination and the perceived fairness of the selection procedure. The dataset contains all 700 completed questionnaires; after the pre-registered exclusions (manipulation check, instructed-response item, speeding), the final analytic sample comprises 581 participants (see S1_sample_flow.csv). Contents: the data in four formats (CSV, Excel, SPSS .sav with value labels, and R .rds); a codebook; the full analysis script (02_full_analysis.R), which reproduces sections 4.1 to 4.8 of the associated article and writes every table to results/; and comparison_with_manuscript.csv, which sets each published figure beside the recomputed one. The README describes the role of every file. Missing data: 172 of the 581 participants (29.6%) left at least one item unanswered, with at most 2.8% missing per item. The README documents how these cases were handled. Software: R >= 4.2 with the packages lavaan, sandwich, car, psych and haven. Licence: CC BY 4.0. Please cite this dataset with its DOI, 10.5281/zenodo.23628806 (see CITATION.cff).