Abstract
This research seeks to develop and test hypotheses about how political disagreement in social networks a ects political behavior. We conduct experimental research to test whether subjects' acquaintances act as independent sources of information, and examine two di erent models of
how such social stimuli may produce eff ects|either via information seeking, or information shortcuts. These tests are important because prior research is ambiguous on whether causal eff ects come from networks, and on potential mechanisms of infuence. Our results back aspects of both models, but more strongly support the notion of disagreement as a heuristic|subjects
primed to consider disagreement before a mock election exhibited a less-orderly information search process; those primed to consider disagreement after the election (but before voting) displayed lower rates of ambivalence, and evidence that such information helped clarify their
decisions.