The paper examines how information on collective action affects the social welfare preferences of voters, in the context of the unfolding COVID-19 crisis. What happens to the social welfare preferences of voters when their expectations concerning collective behavior are met, or even exceeded? And what conversely occurs when these expectations are unmet, and their trust is thereby breached? To explore these questions, we design a quasi-experimental survey directed to a representative sample of the Italian voting age population. In it, a randomly assigned subset of respondents is exposed to real-world information on lockdown compliance rates during the peak of the first Italian COVID-19 lockdown (April 2020). Leveraging this (pre-registered) design, we can then examine the extent to which information on compliance rates affects the social policy preferences of voters, conditional on pre-treatment levels of displayed community trust. We examine voter attitudes toward a broad range of welfare policy dimensions, as our main outcomes of interest: these include social policy generosity, conditionality, and universalism, as well as tax financing and tax progressivity. In addition, we perform several mechanism tests to ascertain the potential causes of uncovered causal effects.