After three decades of scholarship describing why and how students ought to be taught to think historically, this study asks what happens when they are. Ten high school students from a school that incorporated historical thinking into all history coursework repeated the think-aloud task from Wineburg's 1991 study of the cognitive processes underlying the evaluation of historical evidence, reading eight documents with conflicting accounts of the Battle of Lexington. As a cohort, these contemporary students corroborated, sourced, and contextualized more frequently than their 1991 counterparts, despite representing a greater range of overall academic ability. The increase in historical reading did not, however, unambiguously demonstrate a change in their historical thinking. Students tended to source using a binary rating of either reliable or unreliable, corroborate pairs of documents rather than consider how all eight documents in the set created a narrative, and rely upon their ability to recall content information to contextualize. Their performance suggests that they have learned the process of historical thinking without taking up the underlying epistemology. These less sophisticated reading moves raise questions about how well the predominant model for historical thinking in the United States inspires and reflects students' epistemological growth and suggests that there may be a need to revisit how available professional development, educative materials, and research help educators teach historical thinking.