This study examined the relationship between high school exit exams and mathematical proficiency. With the No Child Left Behind (NCLB) Act requiring all students to be proficient in mathematics by 2014, it is imperative that high-stakes assessments accurately evaluate all aspects of student achievement, appropriately set the yardstick by which students will be measured, and clearly communicate these expectations to teachers and administrators. As states across the country transition to the Common Core State Standards, a key goal of this research was to provide current mathematics assessment information for the two consortia charged with the responsibility of creating assessments aligned to these new standards for mathematics. With this goal in mind, the researcher collected assessment data from Massachusetts's, Minnesota's, and Ohio's high school exit exams. These states were selected based upon their use of a comprehensive exit exam to evaluate student readiness for graduation. For each exam, the researcher determined which complexity level, which content strand, and which item format best predicts student proficiency classification. He then critically evaluated each exam based upon its protocol and the National Research Council's definition of mathematical proficiency. Results indicated that no single complexity level, content strand, or item format had the best predictive power for all of the exams. The results indicated that the greatest amount of variation in items' predictive power occurred across complexity level and the least amount of variation occurred across content strand. In addition, the selected exams were found to appropriately assess mathematical proficiency with the following exceptions: (a) they did not meet Norman L. Webb's six-item criterion for categorical concurrence requirements for each complexity-by-content strand category, and (b) two of the four assessments were deemed to have misclassified some students as proficient because the cut score was set too low to require students to earn points from the full range of desirable mathematical behaviors. Results from this study reinforce the idea that exams that intend to assess mathematical proficiency should be designed appropriately and implemented with attention to detail in order to do so. In particular, categorical concurrence and cut score emerged as two critical factors in such assessments. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]