The current study investigates how public attitudes and perceptions about the COVID-19 pandemic evolve over time and influence self-reported health behaviors (e. g., social distancing). Specific attention was paid to respondents' exposure to different news media channels (public vs. commercial). We used data from a two-wave panel study with a 3-week interval (W1 at the start and W2 at the peak of the pandemic) and a large sample of the adult population in Flanders, Belgium (n = 870). The results of mixed ANOVAs indicate that besides a time-effect there was also a significant effect of the different types of news media exposure and respondents' support for protective health measures and behaviors. Whereas, perceived vulnerability to disease, feelings of loneliness, and solidarity were mostly determined by respondents' overall frequency of media exposure, support of governmental measures and self-reported health behaviors were mostly determined by the type of news media exposure. Respondents with a predominantly public/quality news media diet had the highest scores on these variables. A stepwise linear regression analysis with individual's change scores demonstrated that (self-)protective behavior was positively determined by respondents' age, solidarity, and the belief that the measures are necessary, but negatively determined by one's cumulative exposure to commercial/tabloid news media. This longitudinal study provides a new perspective on the role of news media in times of a public health crisis. It offers support for (A) the “double bind hypothesis” (i.e., while news media consumption encourages (self-)isolation, it fosters feelings of loneliness); and (B) the “dual effects hypothesis” (i.e., exposure to commercial/tabloid news media generates different outcomes than exposure to public/quality news media). Affective responses and socio-psychological perceptions are influenced by overall news media exposure, whereas support for the government and its handling of the crisis are mainly determined by one's selection of media channels, whereby audiences of public news media evaluate these outcomes more positively than the audiences of commercial news media channels.