Public decision-makers seek to maximize social welfare, hence to satisfy the population's individual preferences. How do the systems for gathering information on agents' preferences affect the quality and the interpretation of public decision-making? The paper applies a decision-making model to the allocation of a subsidy for public or private transportation. Individual preferences are revealed, at the aggregate level, by individual behaviors in different choice contexts. Using Saari's techniques of probability calculus for voting paradoxes, the loss of information due to these particular conditions of data collection is assessed and analyzed in detail. The author draws conclusions in three areas: information reliability and decision quality; model formulation; and the implementation of welfarism.