AP Computer Science Principles: 2021 Results

The following data reflect the 102,610 students worldwide who took either the paper or the digital AP Computer Science Principles Exam in May. Data from students who tested in June are not yet available.

AP Computer Science Principles score distributions, 2019 vs. 2021

The following table enables comparisons of student performance in 2021 to student performance on the comparable full-length exam prior to the covid-19 pandemic.

AP Score 2019 2021
5 14% 13%
4 21% 22%
3 37% 32%
2 19% 20%
1 9% 13%

Of these 102,610 students, 281 students earned all 100 of 100 points possible across their Create tasks and the end-of-course exam.

The May 17 In-School Paper Exam

The largest exam date for AP Computer Science Principles was May 17, so the following information is specific to the exam version administered on that date.

  • Course Units:
    • AP Computer Science Principles students earned especially high scores on questions about Big Idea 4 (Computer Systems and Networks) and Big Idea 5 (Impact of Computing). A whopping 47% of students earned virtually every point possible in these areas.
    • Student performance across Big Ideas 1–3 was also solid.
    • The most challenging items were algorithm questions with no code. 17% of students earned 0–1 point(s) on these questions, and 8% of students answered all of them correctly.
  • Computational Thinking Practices:
    • Students scored highest on questions related to Practice 5, investigating computing innovations. 30% of students answered virtually all these questions correctly.
    • The most challenging practice for students was Practice 2, developing and implementing algorithms

The Create Task

  • 11% of students earned perfect scores of 6/6 points possible on the Create task.
  • 52% of students earned the point for program purpose and function.
  • 72% earned the point for data abstraction.
  • 38% earned the point for managing complexity.
  • 48% earned the point for procedural abstraction.
  • 44% earned the point for algorithm implementation.
  • 40% earned the point for testing.

The May 26 Digital Exams

To support student access, different testing modes—paper and digital—were essential. To protect exam security, many different exam versions were necessary. Accordingly, to provide students with similar opportunities for success regardless of which version they took, each version of the exam had to be analyzed separately by psychometricians to identify its unique difficulty level so that standards for scores of 3, 4, and 5 could then be separately identified for each exam version. Analyses focused on:

  1. Differences in the testing mode (paper or digital). For sections of the exam that proved easier to take digitally, the digital versions require more points for each AP score. For sections of the exam that proved easier to take on paper, the paper exam requires more points for each AP score.
  2. Differences in the difficulty of specific questions. When exam questions prove easier, more points are required for each AP score, and when exam questions prove more difficult, fewer points are required on one version than another.
  3. The net result for this year’s AP Computer Science Principles exams is that out of 100 points possible, the digital versions were more difficult than the paper version. As a result, depending on the digital version the student received, the student needed to earn 10–14 fewer points than students who took the paper version, to receive an AP score of 3 or higher.