For some subjects, we administer alternate exam versions to small groups of students in the United States on the regularly scheduled testing date. This is to ensure that scores for different exam versions (worldwide) can be compared in a statistically valid way. However, this does also mean that some or all students in one school may take a different version of the exam than the most commonly administered version(s). Overall, in a given year, only a small subset of schools in the U.S. receive and administer more than one version of the exam simultaneously on the regularly scheduled testing date.
These different versions of the AP Exam are distributed to a representative sample of schools of varying sizes across all major U.S. geographic regions. Selected schools are representative of the total population in terms of school characteristics and test performance, and are determined through psychometric and statistical analysis. When a school receives two versions of an exam for a subject, approximately half of the students will take one of the versions given to most U.S. students, and approximately half will take the additional version.