Despite numerous studies document that individuals display a confirmation bias in processing counter-attitudinal information (i.e., they are “motivated reasoners”), there is surprisingly little evidence that these biases lead to attitude polarization. One of the reasons for this apparent puzzle might be that, under certain conditions, liberal and conservatives differ in their tendency to display motivated reasoning and, as a result, to polarize in response to new information. In this study, we argue that motivated reasoning should be asymmetrical when either relational motives to maintain homogenous social networks, or when epistemic motives to reduce uncertainty are triggered. We test these mechanisms directly in a two-wave survey experiment in which we assigned participants to counter-attitudinal evidence on two salient issues in the U.S. - border control and universal healthcare - combined with novel manipulations of relational and epistemic motives. Preliminary findings indicate that in the presence of "social cues" that trigger in-group/out-group motivations, conservatives are more likely to polarize in response to new information than liberals. Besides providing new evidence of how people make up their minds on salient, political issues, this study aims to identify the conditions under which conservatives and Republicans might be more prone than liberals and Democrats to engage in motivated reasoning. By analyzing potential ideological differences, we also aim to shed light on the puzzle of why evidence of attitude polarization seems so elusive, despite a political domain in which “we find bias, bias everywhere” (Taber and Lodge 2016, 82).