This paper examines how a government that controls both public information and pollution emissions can exercise discretion in its choice of pollution signals, and thus influence pollution responses in the population. We develop a model of the government's optimal decision about pollution emissions and information about them, and we apply its results in the context of the information distortions and the adaptation choices of households in Beijing, China. We proceed in three stages. First, we use a simple signal extraction model to motivate why a government may choose to distort information about pollution. Then, we perform a time series analysis on air pollution announcements, to test empirically if the information signal in Beijing is biased when compared to an alternative measure from the US Embassy. This analysis indicates that public information is systematically modified, as predicted by our model. Finally, using data from an original household survey, we examine the effect of the distorted public signal on agents' behavior, and find that those who rely on public media controlled by the government are significantly less responsive to pollution peaks, demonstrating the impact of governmental control over information.