Bayesian network modeling of statistical anxiety, procrastination, and publication motivation in graduate students

Carita A.J.Q.; Cauna E.L.; Panca P.H.; Quispe M.F.M.; Torres-Cruz F. · Journal of Advanced Pharmacy Education and Research (2026) · DOI: 10.51847/gsxyobgehi

Resumen

Research productivity in graduate education depends not only on methodological competence but also on psychological self-regulation. Statistical anxiety, academic procrastination, and publication motivation interact as a dynamic system shaping research development. Conventional correlational methods identify associations but rarely clarify conditional dependency structures. To address this gap, this study developed and validated a Bayesian Network model to estimate probabilistic relationships among these variables in graduate students from a Peruvian public university. Data were obtained from 365 students using validated instruments: the Statistical Anxiety Rating Scale (STARS), the Academic Procrastination Scale, and the Publication Motivation Scale. Bayesian structural learning algorithms identified dependency pathways, and robustness was evaluated through 10-fold cross-validation. Prior statistical training was the strongest predictor of statistical anxiety (Cramér’s V = 0.290, p < 0.001). Students without prior training showed a 55.9% probability of high anxiety versus 25.3% among trained peers. The network positioned statistical anxiety as a central mediating node linking background factors to publication motivation. Structural stability reached 94.3% across validation folds. Simulations showed that reducing anxiety increased the probability of high publication motivation by 133.8%, substantially exceeding the effect of lowering procrastination (42.3%).

Instituciones: Universidad Nacional del Altiplano; Universidad Peruana Unión