Group commitments such as partisanship and religion can bias the way individuals seek information and weigh evidence. This psychological process can lead to distorted views of reality and polarization between opposing social groups. Substantial research confirms the existence and persistence of numerous identity-driven divides in society, but means of attenuating them remain elusive. However, because identity-protective cognition is driven by a need to maintain global and not domain specific integrity, researchers have found that affirming an unrelated core aspect of the self can eliminate the need for ego defense and result in more evenhanded evaluation. This study proposes a competing intervention. Individuals possess numerous social identities that contextually vary in relative prominence; therefore a different means to unbiased cognition may be to make many social identities salient simultaneously, reducing influence of any potentially threatened identity. This may also reduce selective exposure to congenial information, which has not been found with affirmation. This study also advances research on the phenomenon of selective exposure by considering individuals’ interpersonal networks in information search. Because networks are not static, and are instead contextually activated, inducing a more complex representational structure of the self may broaden the set of contacts from whom individuals seek information. The bias-mitigative potential of self-affirmation and social identity complexity is examined here in a series of dispute contexts — two partisan, one religious — over a mining spill, an advanced biofuels mandate, and gene editing technology. Results from the three experiments (total N = 1,257) show modest support for social identity complexity reducing group-alignment of beliefs, behavior, and information search, while affirmation failed to reduce, and in some cases increased, group alignment.