What are the AP Statistics course practices?

The four practices and what each one is worth

The redesigned course, first examined in May 2027, is built on four course practices, each with a published weighting band:

Read as a ranking, that is three large practices and one small one. Analyze Data and Interpret Results share the top band. Collect Data sits one band below. Formulate Questions is the outlier, and by a wide margin: its ceiling of 10% is below every other practice's floor.

The four names are verbs, and they run in the order a real study runs in. You ask a question, you collect data capable of answering it, you analyze what you collected, and you say what the analysis means. The course is not sequenced that way, and it is worth knowing that going in. Topic 1.1 is Introducing Statistics: What Can We Learn from Data?, but Unit 1 then spends topics 1.3 to 1.9 describing and comparing one-variable data before it reaches collection at all, and only ends with four topics on collecting it, including random sampling and experimental design. The practice order is the logic of a study, not the order the course teaches in.

How the four free-response questions map onto them

Section II is where the practices stop describing the course and start being the exam's structure. Three of the four question types are named for the practices they target:

Count the practice labels across the section and there are seven of them, spread like this:

Three of the four practices are named on questions worth 20 of Section II's 40 points, which is 25% of the whole exam each. Practice 1 is named on 10 points, 12.5% of the exam. Those shares overlap rather than add. Question 1 and Question 2 are each counted for two practices and Question 4 for three, while Question 3 is counted for none, which is why the four shares total 87.5% against a section that is worth 50%.

Two patterns fall out of that table that the practice list alone does not show. Practices 3 and 4 are named on exactly the same two questions, so no question type names one of them without the other. And Practice 2 is the only practice named on both the first question and the last one. No question names all four; Question 4 names three, the most of any.

Question 3 is the one that names no practice

Question 3 is the only free-response question whose type names no practice at all. It is labeled by procedure instead: Inference (Hypothesis Test or Confidence Interval). One question out of four is dedicated inference, and the other three are Multi-Focus.

That makes Question 3 the most predictable thing in Section II, and a rehearsed template is worth more there than anywhere else in the exam. But if your Section II plan is that template and nothing else, price it first. Question 3 is 10 of the section's 40 points, 12.5% of the exam. The three Multi-Focus questions are the other 30 points, 37.5% of the exam, three times as much.

None of that makes inference a small part of the course. Units 3 and 4 are both inference units, and between them they carry 25% to 45% of the multiple-choice section. Is inference still on the AP Statistics exam sorts out what the redesign did and did not cut. The point here is narrower: inference is one labeled question out of four in Section II, and the other three are labeled by practice.

Practice 1 is the smallest on every measure available

A band of 5-10% looks like a rounding error next to 25-35%, and three separate signals agree that Practice 1 is the smallest of the four:

1. Its weighting band is the lowest of the four, and its ceiling of 10% sits below the floor of every other practice.
2. It is named on one of the four free-response questions. The other three practices are named on two each.
3. Of the 55 topics in the course, one is titled purely around forming the question, 1.1 Introducing Statistics: What Can We Learn from Data?, and one more shares its title with collection, 1.10 The Investigative Question Revisited and Data Collection, which this page also counts among the four Collect Data topics further down.

That is not permission to skip it. Practice 1 shares Question 1 with Practice 2, which means it is named on the first question in Section II, a question worth the same 10 points as the inference one. A question you cannot get started on costs 12.5% of the exam, an eighth of the whole score, and eats the time you needed for the other three.

What this page can say about Practice 1 is its published name, its band, and the one question type that names it. What a Question 1 task actually puts in front of you is not traceable to a source here, so take the obvious habit it points at, reading what the investigator wants to know and answering that rather than the question you would rather have been asked, as a study suggestion and not as a requirement of the framework.

Analyze and Interpret carry the same band, and the course splits constructing from justifying

Practices 3 and 4 carry the same band, 25-35%, and they are named on the same two free-response questions. It would be easy to treat them as one activity. The course's own topic titles keep building a result and arguing from it in separate topics.

Most of the topic titles in Units 3 and 4 open with a verb, and those verbs fall into four families:

That is 18 of the course's 55 topics, about a third of it, named with one of those four openings.

The pairing matters more than the totals. Each of the five tests in Units 3 and 4 gets both a Setting Up topic and a Carrying Out topic: one proportion, two proportions, chi-square, one mean or mean difference, and two means. And every one of the four Constructing a Confidence Interval topics is immediately followed by a Justifying a Claim topic about that same interval: 3.3 into 3.4, 3.10 into 3.11, 4.2 into 4.3, and 4.7 into 4.8.

Four times out of four, building the interval and justifying a claim from it are separate topics. Reading topic titles is not the same as reading the framework's own alignment of topics to practices, which this page does not have in front of it, so treat that as what the titles show rather than as a quoted mapping. But the shape is hard to miss, and it matches the weighting: interpreting is not the sentence you add after the real work, it carries the same band as the analysis it follows. How to write a hypothesis test conclusion and how to interpret a confidence interval for a proportion are the two templates that earn those points.

Collect Data is heavier than a formula-led revision plan suggests

Practice 2, Collect Data, is banded at 20-30%, one band below the two that tie at the top, and it is named on Question 1 and Question 4, the two ends of the free-response section. Those two questions are 20 of the 40 points.

Unit 1 finishes with four topics on it: 1.10 The Investigative Question Revisited and Data Collection, 1.11 Random Sampling, 1.12 Potential Problems with Sampling, and 1.13 Experimental Design.

Those four are among the least numerical topics on the course, which is exactly what makes them easy to drop from a revision plan built around formulas. They also carry the answer to a question the exam keeps asking: what are you allowed to conclude from data collected this way. Random selection is what lets you generalize to a population. Random assignment is what lets you claim a cause. One without the other limits what you may write, whatever the p-value says.

Drill it with experiments vs observational studies, can you generalize results, and scope of inference practice.

What the labels do not tell you, and how to drill anyway

Three things a reader arrives here wanting, stated plainly rather than guessed at.

The labels do not give a point split. Question 1 is Multi-Focus on Practices 1 and 2 and it is worth 10 points. The label names the practices; it does not say how the 10 points divide between them. This page does not invent a division, and you should be suspicious of any that quotes one.

This page does not attach the weighting bands to a section. The bands are published per practice. Turning them into a count of multiple-choice questions, or into free-response points, would need a source this page does not have. What they give you is a ranking of emphasis across the course, which is enough to plan with.

There is no published score conversion for this exam. It has not been administered yet, so nothing about its cut points has been released. What is a good AP Statistics score covers what can and cannot honestly be said about that.

A study plan that follows the practices rather than fighting them: rehearse the inference template for Question 3, then give at least as much time to the three Multi-Focus questions, which are 30 points against its 10. Practice describing what a design does and does not license for Question 1, reading and comparing distributions for the analysis half, and writing conclusions in context for the interpretation half. How to pace the AP Statistics exam has the timing, the FRQ guide has the templates, the FRQ format page has the section in full, and how to study for AP Statistics turns it all into a schedule.