Objective This article examines the distinction between group‐based issue opinion formation (what we term “following the crowd”) and idiosyncratic or nongroup‐based formation (what we term “thinking outside of the box”). The argument put forth is that issue saliency can lead citizens to think about issues in nongroup‐based terms. Method We use heteroskedastic regression to measure the degree to which group‐based variables explain issue opinions. Using group variables (demographics, party identification, etc.) to estimate respondents' issue responses means that nongroup variation is soaked up by the error term. Results We find that citizens who view an issue as highly salient are more likely to “think outside the box,” while citizens who view an issue as less salient are more likely to “follow the crowd” by defaulting to their group memberships and identifications. Conclusion Our results indicate that response variability (less consistency within groups) on issue opinions is not always the result of uncertain citizens, nonattitudes, or measurement error. In some situations, greater response variability can reflect a deliberative and policy‐based form of opinion formation.