Multiple-choice questions (MCQs) are commonly used in undergraduate introductory science, technology, engineering, and mathematics (STEM) courses, and substantial evidence supports the use of student-created questions to promote learning. However, research on student-created MCQ exams as an assessment method is more limited, and no studies have investigated whether student-created exams address inequities in STEM outcomes. The current study explored student-created and instructor-created MCQ exams with students in an undergraduate human physiology course (N = 46). Descriptive statistics, Pearson correlations, and a paired samples "t" test compared student performance on the two versions. Multiple methods assessed exam quality, including the percentage of students who responded correctly to each question (difficulty), the corrected item-total correlation coefficient (discrimination), and an objective rater score (cognitive complexity). A series of four repeated measures factorial analyses of variance examined demographic subgroup performance differences. Students performed significantly better on student- rather than instructor-created exams. Both versions discriminated similarly and were moderately to strongly correlated with each other. However, student-created exams had a larger percentage of "easy" questions. Students who identified as first generation and/or low income performed significantly higher on student-created exams, but still failed to achieve the same level of performance as their peers. Student-created MCQ exams may serve an important role as part of a comprehensive assessment plan.