Assessing students' historical understanding has been a long-standing challenge in history education. One of the widely used tools for accomplishing this task is the large-scale standardized test, the results of which are used as an indicator of student knowledge and skills in the social sciences/history. At the national level, the National Assessment of Educational Progress (NAEP), often referred to as the "nation's report card," is considered as the provider of a "common yardstick" by which to compare student knowledge and skills nationwide in social sciences/history along with many other content areas. Despite their usefulness to educators and the public alike, large-scale assessments in the form of standardized testing at state and national levels have been criticized for addressing fragmented factual knowledge rather than students' historical and inquiry skills. In search of ways to address the problem of focusing on fact-based questions, it was Document-Based Questions (DBQs) that these large-scale assessments have incorporated. Using historical knowledge, students are expected to interpret the meaning of primary sources (e.g., speeches, letters, diaries, quantitative materials, political cartoons, photos, and paintings) and relate the meaning of the sources to the historical context, given the students' historical understanding. Thus, ideally, DBQs require students to exhibit knowledge of history while demonstrating critical analysis skills. The question then, is whether the NAEP uses DBQs in a way that calls for students to do more than just regurgitate historical facts. To answer this question, this study explores how the NAEP uses primary sources--specifically visual images--to assess eighth and twelfth grade U.S. history students' historical knowledge and thinking skills. Though scores on the NAEP have traditionally served as a barometer for both researchers and the public to judge the quality of teaching and learning in U.S. history classrooms, very few studies have been conducted regarding the questions themselves. In this study, the authors neither idealize the NAEP assessment nor criticize it. They use it as a case for studying the use of historical images in assessing students' historical thinking skills on large-scale assessments, and for exploring the possibilities and challenges of using visual images to teach students how to think historically. In this regard, the authors sought to answer the following questions regarding the U.S. history section of the NAEP: (1) How are visual images used as primary sources in the NAEP assessment?; and (2) To what degree are visual images used to assess students' historical thinking skills?