AP Statistics

AP Statistics FRQ Tips

The free-response section of the AP Statistics exam is worth 50% of your total score — more than any other AP exam's FRQ section. You'll have 100 minutes for 6 questions (5 short + 1 investigative task). Success depends on clear communication, proper statistical vocabulary, and always writing in context.

Key Takeaways

FRQs are 50% of your AP Stats score — the highest weight of any AP exam. You cannot score well without strong FRQ skills.
Always write in context. "There is evidence of a difference" is incomplete. "There is convincing evidence that the mean battery life of Brand A is greater than Brand B" earns the point.
Follow the four-step process for every inference problem: (1) State hypotheses/parameter, (2) Check conditions (random, normal, independent), (3) Calculate, (4) Conclude in context.
Show your conditions with numbers, not just words. Don't write "Normal condition met." Write "np̂ = 42 ≥ 10 and n(1−p̂) = 58 ≥ 10, so the sampling distribution is approximately normal."

How to Write AP Stats FRQs That Score Well

The AP Statistics FRQ section is unique among AP exams because it's worth a full 50% of your score and it's almost entirely about communication. The math calculations are secondary — your TI-84 does most of the work. What the graders are evaluating is whether you can explain statistical concepts clearly, in context, using proper vocabulary.

The 6 FRQs break down as follows: Questions 1-5 are shorter, targeted questions (about 12-13 minutes each), and Question 6 is the investigative task (about 25 minutes). The investigative task is longer and often combines multiple statistical concepts into one scenario.

For inference questions (which appear on almost every exam), the four-step process is your template. Never skip a step. Step 1: State your hypotheses using proper notation (H₀: μ = 50, Hₐ: μ > 50) and name what μ represents in context ("where μ is the true mean weight of cereal boxes produced at this factory"). Step 2: Name the procedure (one-sample t-test) and check all conditions — this is where most students lose points, so be thorough. Step 3: Show the test statistic and p-value. Step 4: Write a conclusion that connects the p-value to your decision and states the conclusion in context.

For exploring data questions (describing distributions, comparing groups), use the framework: shape, center, spread, unusual features. Always compare using comparative language when two groups are involved. "The median for Group A (42) is higher than the median for Group B (35)" — not just "Group A's median is 42."

For experimental design questions, you need to describe: random assignment to treatment groups, the treatments being compared, the response variable, and how you'll compare results. Mention control/placebo groups and blinding if relevant.

The biggest predictor of FRQ success is practice with real rubrics. Do a practice FRQ, then grade yourself using the College Board rubric. You'll quickly learn exactly what language earns points and what doesn't.

Maximizing Points on Every Question

Specific language matters. Here are the phrases that earn points and the phrases that don't:

For conclusions of hypothesis tests: Say "There is [not] convincing evidence that [specific claim in context]" — NOT "We accept/prove the null hypothesis." You never "accept" H₀ or "prove" anything in statistics.

For confidence interval interpretations: Say "We are 95% confident that the true population mean... is between... and..." — NOT "There is a 95% probability that the true mean is in this interval." The true mean is fixed; the interval is what varies from sample to sample.

For p-value interpretations: Say "The probability of observing a result as extreme as or more extreme than the sample result, assuming H₀ is true, is 0.03" — NOT "The probability that H₀ is true is 0.03."

For the investigative task (Q6): Don't let its length intimidate you. Read the entire question first, then work through each sub-part. It's broken into manageable pieces. Partial credit is available for each part, so answer what you can even if you get stuck on one sub-part.

Time management: aim for 12-13 minutes per short question and 25 minutes for the investigative task. If a question has you stuck after 5 minutes, write what you know, move on, and come back later.

Common Mistakes Students Make

Forgetting to write in context. This is the #1 point-loser on AP Stats FRQs. Every answer should reference the specific scenario — the type of data, the population, the variable being measured. Generic statistical statements without context receive reduced credit.
Saying "accept H₀" instead of "fail to reject H₀." In statistics, we never accept the null hypothesis — we only fail to find sufficient evidence to reject it. This distinction matters and costs points every year.
Checking conditions without showing the math. "The normal condition is met" is not enough. You must show: for proportions, that np̂ ≥ 10 and n(1−p̂) ≥ 10 (with actual numbers); for means, that n ≥ 30 or the population is approximately normal.
Confusing association with causation. If the data comes from an observational study, you cannot conclude causation — only association. Only randomized experiments support causal conclusions. The AP exam tests this distinction explicitly.

Frequently Asked Questions

How much is the FRQ section worth on AP Statistics?

The FRQ section is worth 50% of your total AP Statistics score — split evenly with the 40 multiple-choice questions. This is the highest FRQ weight of any AP exam, making FRQ practice especially important for AP Stats.

What is the investigative task on the AP Stats exam?

Question 6 is the investigative task — a longer, multi-part question that often combines data analysis, probability, and inference into one scenario. It requires more reading and thinking than the shorter FRQs but is worth the same number of points. It often presents an unfamiliar context and asks you to apply statistical reasoning in a novel way.

How do I check conditions on AP Stats FRQs?

For every inference procedure, check three conditions: (1) Random — was the data collected using a random method? (2) Normal — is the sampling distribution approximately normal? For proportions: np̂ ≥ 10 and n(1−p̂) ≥ 10. For means: n ≥ 30 (CLT) or population is approximately normal. (3) Independent — are observations independent? If sampling without replacement, check that n < 10% of the population.