Abstract Moral judgements are often treated as stable preferences perturbed by noise. Quantum cognition, however, models decision-making as probabilistic superpositions that collapse under contextual measurement. We reanalysed responses from 583 Brazilians evaluating the Age of Penal Majority (APM) across three framings—online juridical, in-court legal and public-square civic. Participant moral states were represented in a three-dimensional Hilbert space via principal component analysis (PCA), with context-specific measurement bases yielding model-implied counterfactual reversal probabilities of 0.50–0.76 at the population level. These closely matched Bayesian hierarchical estimates and the quantum representation achieved slightly better widely applicable information criterion (WAIC) and leave-one-out (LOO) cross-validation scores than a propensity-weighted logistic baseline. Propensity-weighted bootstrap and hierarchical Bayesian analyses addressed sample imbalance and yielded convergent reversal estimates. The results extend quantum-inspired modelling to a real, socially consequential dilemma and show that legal and policy decision-support frameworks must account for context-driven shifts in public moral judgement rather than assume fixed preferences.
Caldas et al. (Wed,) studied this question.