Democracies assume accurate knowledge by the populace, but the human attraction to fake and untrustworthy news poses a serious problem for healthy democratic functioning. This is exacerbated by the fact that political groups are more divided along ideological lines than at any point in decades. Although partisan antipathy accounts for deep rifts on a number of important issues, scientists still know very little about the neurobiological roots of polarization. In this talk, I will articulate why and how identification with political parties—known as partisanship—can bias information processing in the human brain. My research suggests that the roots of political partisanship are basic representations of group membership — grounding political differences in basic tribal distinctions between us and them. I will introduce a value based model of belief for understanding the influence of partisanship on these cognitive processes. This model bridges politics, psychology, and neuroscience to help explain why people place party loyalty over policy, and even truth. Specifically, the model explains how partisanship has cognitive consequences that extend well beyond motivated political reasoning, to memory, implicit evaluation, and even perceptual judgments. In the final section, I will discuss strategies for de-biasing information processing to help create a shared reality across partisan divides.