The impact of mass media on political outcomes like government accountability (Besley and Prat, 2006), voting behavior (Dellavigna and Kaplan, 2007), and public policy (Strömberg, 2004; Eisensee and Strömberg, 2007) has been well documented. However, as widely appreciated as the forceful role of media in politics is, its troubling tendency toward bias is also widely acknowledged. Existing research demonstrates the origins of media bias from both demand (Strömberg, 2004; Gentzkow and Shapiro, 2010) and supply (Baron, 2006; Druckman and Parkin, 2005) sides. However, most previous studies have focused on the bias of individual outlets. In contrast, this paper argues that not only is each individual outlet biased in its reporting of political issues, but also that multiple outlets tend to similarly report on the same issue. In other words, there exists a possible collective bias in the coverage of political issues. Moreover, this collective bias may result in a collective failure, in which information regarding the superior policy is excluded from the news, the inferior policy ultimately gets implemented, and, in this way, overall social welfare is harmed. The notion of collective bias is based on the following observations. (I) The mass media’s coverage of political issues is likely to be affected by partisan individuals’ preferences, but could also affect the decisions of non-partisan individuals, those who have no exante bias toward any issue. Gentzkow and Shapiro (2006) argue that newspapers cater to audiences’ preferences, because audiences are more likely to regard an outlet that reports ‘‘like-minded news’’ as high quality. In a later study, Gentzkow and Shapiro (2010) analyze US daily newspapers to prove that the mass media significantly responds to consumers’ views in their reporting. The assumption of selective exposure to like-minded news is in line with studies in social psychology, specifically those of cognitive dissonance. In contrast, non-partisan individuals lack ex-ante bias and, in turn, selective exposure. In other words, the mass media could directly affect the private actions and/or political decisions of non-partisans by manipulating the information to which they have access. In this paper, we focus on the role of non-partisan instead of partisan individuals, unlike most studies in political economics. We did this for three reasons. First, it is the non-partisans who finally decide the outcome of elections. Second, all partisans, prior to choosing partisanship, were, at one point, non-partisan. Third, most young people who have a decisive impact on their country’s future political direction are or have been non-partisan. Studies on the influence of mass media on non-partisans help us to understand the dynamics of political change in a society. (II) The expanded use of social media, such as Facebook and Twitter, has intensified the exchange of information among individuals. Nowadays, audiences obtain information from mass as well as social media because informed audiences post content on social media. Meanwhile, audiences, especially younger ones, have gradually shifted from purchasing traditional print media products to reading news online, typically by downloading apps and then reading content electronically on mobile terminals.