AP Computer Science Principles Create PT

AP Computer Science Principles is the College Board's broad, concept-first computer science course. It asks candidates to reason about data, the internet, algorithms, programming, and the societal impact of computing, and it does so through an unusual two-part exam: a 70-minute multiple-choice section and two sustained performance tasks submitted digitally through the College Board's portal. For a student choosing between AP Computer Science Principles and the more code-heavy AP Computer Science A, the deciding question is rarely difficulty. It is format. AP CSP rewards clear written explanations, defensible design decisions, and reflective analysis of computing innovations; AP CSA rewards syntactically precise Java code on paper.

This article dissects the AP Computer Science Principles Create Performance Task, the rubric rows that separate a 3 from a 4, the algorithm point that most candidates drop, and the way the multiple-choice exam and the Explore Performance Task are scored. The goal is operational: by the end, a reader should know which rows matter, what language to use, and how to budget 12 weeks of preparation around the two digital submissions.

What the AP Computer Science Principles assessment actually measures

AP Computer Science Principles is built around five Big Ideas, sometimes abbreviated as CRD: Creative Development, Data, Algorithms and Programming, Computer Systems and Networks, and Impact of Computing. The two performance tasks and the 70 multiple-choice items together sample all five. The course does not require a specific language; the College Board permits a list of approved languages and IDEs, and the Create PT submission must include both a program written by the student and a written response describing design choices. For most candidates reading this for the first time, the first surprise is that AP CSP is roughly 40 percent writing. The multiple-choice section contains about 70 standalone and set-based questions; the Create PT contributes a written response of around 400 words plus code excerpts; the Explore PT is purely written, anchored on a student-chosen computing innovation.

The score is composite. The multiple-choice section and the two performance tasks are combined into a single AP score on the standard 1 to 5 scale. Historically the Create PT has accounted for the largest single share of total points, which is why nearly every tutoring conversation about AP Computer Science Principles eventually returns to the question: which rows in the Create PT rubric actually decide a 5, and which rows silently push a candidate from a 4 down to a 3?

A practical orientation: the Create PT is scored against six rows, the Explore PT against five rows, and the multiple-choice section against a fixed set of items. Each row carries a small number of points, but the rows are not equally generous. The next two sections walk through both performance tasks in the order candidates encounter them.

Inside the Create Performance Task: the six rows that decide a 3, 4, or 5

The Create PT asks a candidate to design and implement a program of their own choice, then submit a video, the program's code, and a written response. The College Board scores the task against a published rubric of six rows. Each row awards a point when the submission meets the row's criteria. There is no partial credit within a row; a row is either earned or it is not. This is the most important structural fact about AP Computer Science Principles scoring, and it is the reason two candidates with similar overall submissions can finish a full point apart.

The six rows, in the order they appear on the rubric, are:

• Program purpose and development row: the program must have a clear, student-articulated purpose, and the video must show the program running with input and output that demonstrate that purpose.
• Algorithm row: the written response must describe an algorithm used in the program in enough detail that the algorithm could be re-implemented.
• Abstraction row: the response must identify and explain at least one list or collection used to manage complexity, including what is stored and how the choice of list simplifies the program.
• Code segment row: the response must include a code segment that implements the algorithm described in the algorithm row.
• Testing row: the response must describe two distinct calls to the selected procedure, each with different arguments, and describe what each call demonstrates.
• Reflection row: the response must describe one way the program could be improved, naming a specific change and the effect of that change on functionality.

For most candidates, four of these rows are recoverable with a careful rewrite during the final two weeks. Two of them, the algorithm row and the testing row, are the rows that get dropped in a way that cannot be patched with a one-line edit. The next section examines the algorithm row in detail because it is the single most mis-scored row in the Create PT.

The algorithm row: where most AP Computer Science Principles candidates lose a point they can recover

The algorithm row awards its point when the written response describes an algorithm used in the program, identifies the algorithm's name or its core strategy, and includes enough detail that the algorithm could be re-implemented from the description alone. Note three things. First, the description must be of an algorithm, not of a single statement or of an entire program. Second, the description must be specific to the candidate's program, not a textbook paraphrase. Third, the description must be sufficient on its own; the reader should not need to consult the code to understand the algorithm.

The classic failure pattern is a candidate who describes what the program does rather than how the algorithm works. Consider a program that sorts a list of student names alphabetically. A common student response says, "The program uses a sort algorithm to put the names in alphabetical order." That sentence describes the program's behaviour, not the algorithm. A scoring response says something like, "The program uses a selection sort that iterates through the list, locates the smallest remaining name, and swaps it into the next unsorted position, repeating until the list is sorted." The difference is operational detail, not vocabulary.

A second failure pattern is a candidate who describes the algorithm in terms of language-specific syntax. "The program uses a for loop with i in range(len(names)) and an inner for loop that compares names[j] with names[j+1]" is a description of code, not of an algorithm. The algorithm row asks for the algorithm, in language that a reader could translate into any of the approved AP CSP languages. For most candidates I have seen, switching from syntax-level to algorithm-level language is the highest-leverage rewrite available during revision.

A third failure pattern is a candidate who describes the algorithm in only two sentences. The rubric does not impose a word count, but the description must be sufficient. Two sentences is almost never sufficient unless the algorithm is exceptionally simple. A useful self-check: if a classmate were to read the algorithm description and try to implement the algorithm from that description alone, would they be able to? If the answer is no, the description is too thin.

A practical exercise: take any program written for the Create PT, identify the most interesting procedure or function, and try to write a 100-to-150-word description of the algorithm it implements, free of language-specific syntax. If the description cannot reach 100 words without padding, the procedure is probably not the right one to feature in the algorithm row, and a more substantive procedure should be chosen.

The Explore Performance Task: how the rubric scores the computing innovation and its impact

The Explore PT asks a candidate to choose a computing innovation of personal interest, research it, and submit a written response of about 300 words and an artifact. The artifact is typically a visual or graphical element such as a concept map, a timeline, or a labelled diagram. The Explore PT is scored against five rows:

• Innovation row: the response must identify a computing innovation and explain what it is.
• Use row: the response must describe the innovation's intended purpose or function.
• Data row: the response must identify the data the innovation uses, the source of that data, and how the innovation processes it.
• Effects row: the response must describe a beneficial and a harmful effect, or a single effect from both a beneficial and a harmful perspective.
• Evidence row: the response must cite at least one source, with a citation that is specific enough to be located.

The Explore PT is a less common source of point loss than the Create PT, but the effects row is a frequent silent-killer. Many candidates describe one effect in favourable terms and one effect in unfavourable terms without specifying the stakeholder or context. A scoring answer names the stakeholder, names the mechanism, and explains the causal chain. "Social media can be harmful because it spreads misinformation" is descriptive; "Social media platforms that use algorithmic recommendation have increased the visibility of low-credibility news content for users aged 18 to 29, contributing to measurable shifts in political polarisation" is causal. The difference is mechanism, not length.

The data row is the second-most-dropped row. Candidates often describe what the innovation does, not what data it consumes. A scoring data row names a specific category of data, names a specific source or class of sources, and explains one transformation the innovation performs on that data. The College Board is explicit that the data row is satisfied by explanation, not by code; the Explore PT contains no code at all.

Big Idea 1 to Big Idea 5: the unit map and where marks cluster on the MCQ

The 70-item multiple-choice exam is organised around the five Big Ideas. The College Board's course description publishes approximate weightings, and the practical implication for preparation is that some units are denser in point-yield than others. A summary, in plain language:

Two of the five Big Ideas dominate the multiple-choice section. Big Idea 3, Algorithms and Programming, is where most candidates can recover the largest number of points through deliberate practice, because the question types are predictable: trace a code segment, predict the output of a procedure given specific inputs, identify the bug, or choose a working implementation. Big Idea 2, Data, is the second high-yield area, with a focus on bit-level calculations and the difference between lossy and lossless compression.

For a 12-week study plan, a defensible split dedicates roughly 35 percent of weekly time to Big Idea 3, 20 percent to Big Idea 2, 15 percent to Big Idea 4, 15 percent to Big Idea 5, and 15 percent to Big Idea 1 plus performance-task writing.

The 70-minute multiple-choice exam: question types, set-question logic, and pacing

The multiple-choice section of AP Computer Science Principles contains two formats: roughly 55 to 60 standalone multiple-choice items and two question sets, each containing a stimulus plus four to six related items. The exam is 70 minutes total, with no separate time budget for the two sections. In practice, candidates who finish the standalone items in about 40 minutes leave themselves around 30 minutes for the two sets, which is the right shape for a candidate aiming at a 4 or 5.

The standalone items test recall and short reasoning. A typical item shows a short code segment and asks for the output, or shows a scenario and asks which Big Idea it best illustrates. The set questions test depth. A typical set presents a 15-to-25 line code excerpt, or a description of a computing system, and asks a sequence of related questions that range from reading comprehension to algorithm design.

Three pacing rules serve most candidates well. First, do not spend more than 45 seconds on any standalone item. Second, on a set question, read the stimulus completely before looking at the items, then answer the items in order, because they are typically written so that earlier items are easier. Third, when two answer choices are both defensible, choose the one that names a specific mechanism, since the College Board writes distractors at the level of description rather than mechanism.

The set-question format is also where Big Idea 4, Computer Systems and Networks, is most often tested. A typical set presents a description of a multi-tier network with redundant paths and asks which design property provides fault tolerance, which tier is the single point of failure, and how adding parallel connections would affect throughput. Candidates who can name the property, point at the diagram, and explain the mechanism recover the full set of points.

Building a 12-week AP Computer Science Principles study plan around the two performance tasks

Because the Create PT and Explore PT are digital submissions with their own deadlines, the 12-week plan should anchor on the submission dates, not on the multiple-choice exam date. A practical structure:

• Weeks 1 to 3: Big Idea 1 and the Explore PT innovation. Read the published Explore PT rubric in full, choose a computing innovation the candidate finds genuinely interesting, and begin a research log. By the end of week 3, the candidate should have a working thesis for the innovation row, the use row, and the data row.
• Weeks 4 to 7: Create PT design and implementation. Choose a program that uses at least one list, a procedure with parameters, and a non-trivial algorithm. Build the program incrementally, with weekly checkpoints. By the end of week 7, the program should be feature-complete, with a draft of the video walkthrough.
• Weeks 8 to 9: Create PT written response. Draft the algorithm row, abstraction row, code segment row, testing row, and reflection row against the rubric. Have a teacher or tutor score each row against the published criteria. Rewrite any row that is not yet at scoring level.
• Week 10: Explore PT final submission. Finalise the Explore PT written response, the artifact, and the citations. Run a second pass against the rubric.
• Week 11: Multiple-choice review. Spend five sessions on Big Idea 3, three on Big Idea 2, two on Big Idea 4, two on Big Idea 5, and one on full-length practice items. Time each session at 70 minutes to build exam stamina.
• Week 12: Final review and simulation. One full-length multiple-choice simulation under timed conditions. One final pass of the Create PT and Explore PT for transcription errors and missing rows.

For a candidate whose Create PT submission is borderline, the single highest-value activity is rewriting the algorithm row against a teacher-marked copy of the rubric. The next highest-value activity is rewriting the testing row so that the two calls described in the response match the program in the video.

Common pitfalls and how to avoid them on AP Computer Science Principles

The Create PT and Explore PT share a small number of recurring failure modes. None of them is hard to fix once spotted, but each is a full-row point loss when left in place.

• The algorithm row written as a description of the program. The fix is to identify the procedure to be featured, write the algorithm description in 100 to 150 words, and ask whether the description is sufficient on its own.
• The testing row with two calls that have the same arguments. The two calls must have different arguments and must demonstrate different behaviours. A scoring testing row names the arguments, names the expected output, and explains what each call demonstrates.
• The abstraction row that names a list without explaining its role. "The program uses a list" is descriptive; "The program uses a list to store all the items the user has entered, which lets the search procedure iterate over the full input without re-querying the user" is explanatory.
• The Explore PT effects row that names one effect from one perspective. The effects row requires a beneficial and a harmful effect, or a single effect from both perspectives, with a named stakeholder and a mechanism.
• The Explore PT evidence row with a vague citation. "According to a study" is not a citation. "According to a 2020 study by the Pew Research Center on social media use among U.S. adults" is a citation, even without a URL.
• The multiple-choice exam with no time budget for the set questions. The set questions carry more points per minute than the standalone items. A candidate who runs out of time on the set questions is leaving more points on the table than a candidate who runs out of time on the standalone items.

Two additional pieces of advice that are easy to over-look. First, the Create PT video is a component, not a separate submission; the video must show the program running with input and output, and it must be uploaded in a format the College Board accepts. A candidate whose video shows only the code being typed, or only the source code with no execution, has not met the program purpose and development row. Second, the College Board forbids collaboration on the program code; two candidates with the same algorithm description and the same testing calls will both lose points, and both schools will be notified.

Where to push next on AP Computer Science Principles

For most candidates, the gap between a 3 and a 4 on AP Computer Science Principles is not knowledge; it is language. The algorithm row, the abstraction row, and the testing row all reward specific, mechanism-level description rather than general praise of the program. The next concrete step is to take the current Create PT draft, score it row-by-row against the College Board's published rubric, and rewrite the two lowest-scoring rows. AP Courses' one-to-one AP Computer Science Principles Create PT programme does exactly that: it analyses each candidate's Create PT draft against the six rubric rows, identifies the two rows with the highest recovery value, and turns a 3-or-4 target into a concrete row-by-row revision plan.

Frequently asked questions