Abstract
<p>A key puzzle in the political information literature is why factual information reliably updates citizens’ beliefs about the fact in question but rarely shifts downstream political attitudes. Using two preregistered survey experiments (N ≈ 3,250) covering eleven issues across economic indicators and misstatements by prominent political figures, this paper makes two contributions. First, we test whether the muted effect of factual information on downstream attitudes holds for outcomes that are less entrenched and more proximate to the information thanthose tested in prior work—such as economic evaluations and perceptions of politicians’ honesty and competence—and find that even these outcomes respond only modestly. Second, we develop and empirically test a framework of six adjacent belief dimensions—regarding causes,relevance, and intent—through which factual information may be reinterpreted in ways that reshape its political meaning. We find evidence of adjacent belief updating, although concentrated on specific dimensions rather than widespread. The strongest finding is that individualswho learn that a favored politician made an inaccurate claim concede the inaccuracy but shift toward viewing the statement as figurative, with sizable effects across two independent studies. Importantly, where adjacent belief adjustment occurs, its direction is contingent on individuals’ predispositions. Uncongenial information triggers adjustment that attenuates the political implications of the fact, while congenial information triggers no such adjustment or small shifts in the opposite direction. Together, the results illustrate how parallel trends in factual accuracy can coexist with divergent trends in how those facts are politically interpreted.</p>