A study explored the way various levels of narrative voice in historical texts influence high school students as they carry out a common historical problem-solving task. Participants, 30 students in a suburban Seattle (Washington) high school, wrote a historical essay after reading an introductory text excerpt and six historical documents. Half of the students began the task by reading an excerpt from a popular high school world history textbook with an anonymous author writing in the third person and revealing little about personal opinions, perspectives, or information about historical epistemology. The other half began by reading a text that had the same primary historical information but featured a visible author who wrote in the first person and revealing beliefs, perspectives and information regarding historical epistemology. A pretest gauged students' prior knowledge of Caesar and his assassination, and six were selected to think-aloud through the entire task. All participants used a packet of historical documents examining Caesar's murder. The introductory text that framed the task was different for each group, while the problem-solving task was identical. For the majority of those in the anonymous author group, it appears the discourse schema brought to bear on this task was a straightforward "school writing" mode of communication. Essays produced by the visible author group were about 25% longer than those of the other group. In summing up think-aloud statements, it became clear that participants in the visible author group were far more likely to establish relationships with text authors, hold mental conversations with and about them, and give more thought to the primary historical information embedded in those texts. Appended are text excerpts, historical documents, and scoring rubrics. (Contains 9 tables and a 51-item bibliography.) (BT)