Open Science Behaviour Dataset with Synthetic Data and Monte Carlo Validation

This dataset supports a study on modelling open science behaviour using a bootstrap-based Monte Carlo approach. It includes a cleaned dataset derived from a survey of researchers affiliated with Spanish institutions, a synthetic dataset generated through bootstrap resampling, and comparative statistical outputs validating the structural consistency between real and synthetic data. The dataset focuses on behavioural dimensions related to open science practices, including data deposit, data reuse, transparency attitudes, peer review perceptions, and conflict perception. A composite index measuring overall open science attitude is also included. Version 2.0 accompanies the peer-reviewed camera-ready release of the associated paper. It extends v1.0 by increasing the synthetic sample size (M = 500), adding B = 1000 bootstrap replicates for 95% confidence intervals, reporting Cronbach's α for the composite index (α = 0.675), and providing the four statistical tables reported in the paper: descriptive statistics, Kolmogorov–Smirnov and chi-square fidelity tests, data-deposit category frequencies, and Spearman correlation matrices (real vs synthetic). Substantive findings are unchanged. This resource may support further research in information behaviour, open science adoption, behavioural modelling, reproducibility, and synthetic data generation.