Cognitive validity examines the relationship between what an assessment aims to measure and what it actually elicits from test takers. The present study examined whether multiple-choice items from the National Assessment of Educational Progress (NAEP) grade 12 U.S. history exam elicited the historical thinking processes they were designed to measure. Think-aloud data from 27 accomplished high school students revealed that in no instances did students engage in the intended processes. Rather, the items typically elicited three construct-irrelevant processes: factual recall/recognition, reading comprehension, and test-taking strategies. Further, findings revealed that although the items often prompted students to engage in factual recall, they were often not sound indicators of student knowledge. Implications for history testing and alternative forms of assessment are discussed.