This study examined whether and how (in)civility and the presence of supporting evidence in disagreeing comments influence individuals’ attitude polarization. The study used a 2 (civility vs. incivility) × ۲ (evidence vs. no evidence) factorial design involving reading dissimilar viewpoints in Facebook comments. The results showed that exposure to uncivil opposing comments, compared to exposure to civil disagreeing comments, led to lower levels of willingness to read more comments and greater levels of negative emotions and attitude polarization. However, the presence or absence of supporting evidence in comments did not have any significant effect on the outcome variables. The findings suggest that it is the civility or incivility of information that influences whether exposure to dissimilar perspectives either mitigates or reinforces individuals’ attitude polarization. This study also suggested willingness to read more comments and negative emotions as two mediating factors between exposure to uncivil/civil disagreeing comments and attitude polarization.

Political polarization is a growing concern in many countries and has drawn scholarly attention to discover the factors associated with it (Gramlich, 2017; Tsfati & Chotiner, 2016). As evidence shows that individuals’ exposure to like-minded perspectives is significantly related to political polarization (Garrett, 2009; Stroud, 2011), examining what could reduce political polarization has become important. Deliberative democratic theorists and empirical studies have suggested that exposure to diverse perspectives plays a role in increasing understanding of the opposing side, which may in turn reduce extreme attitudes (Huckfeldt, Johnson, & Sprague, 2002; Huckfeldt, Mendez, & Osborn, 2004). However, exposure to diverse or dissimilar information does not always work in the way deliberative democratic theorists have expected. Their expectation is that exposure to different perspectives makes individuals consider contrasting viewpoints, which leads them to understand the other side and develop greater levels of political tolerance; in this way, people’s attitude polarization is reduced (Huckfeldt, Johnson, & Sprague, 2004; McPhee, Smith, & Ferguson, 1963; Mutz, 2002). On the other hand, some studies have shown that exposure to dissimilar views amplifies individuals’ preexisting beliefs and produces more extreme attitudes, rather than mitigates them (Taber & Lodge, 2006). The latter argument is attributed to biased information processing or motivated skepticism, by which individuals tend to give more weight to information that supports their own position and reinforces their viewpoints and consequently to scrutinize or counter-argue dissimilar information to protect their views (Ditto & Lopez, 1992; Taber & Lodge, 2006).