AP Statistics
AP Statistics FRQ Tips: How to Write Answers That Actually Score
days until your AP Stats exam
Thu, May 7 · Afternoon session
How the AP Stats FRQ points are scored
Five short FRQs (~4 points each) and one Investigative Task (~8 points). The Investigative Task asks you to apply statistical reasoning to a novel situation - it's the highest-leverage FRQ to practice.
Points per FRQ type
Suggested time budget · 90 min
Short FRQs (×5)
20 pt · 55 minStandard inference, probability, and study design prompts. 1–3 sub-parts each.
Investigative Task
8 pt · 35 minNovel application of statistical reasoning. Open-ended. Highest-leverage prep target.
Exam composite weighting
6 FRQs carry the same weight as 40 MCQs
Decode the task word before you write
Every AP FRQ begins with a task word. It tells you the minimum sentence type required for the point. Writing too much costs time; writing the wrong kind costs the point outright.
Want to know if you're actually doing this?
Write one Stats FRQ. Get your setup, conditions, and interpretation scored line by line.
Weak vs. strong: hypothesis test
Inference FRQs are scored on four steps: hypotheses, conditions, mechanics, and conclusion. Missing any one step caps your score.
Prompt
A researcher claims that more than 60% of adults exercise weekly. A random sample of 200 adults found 130 who exercise weekly. Test at α = 0.05.
Weak answer
"p-hat = 130/200 = 0.65. z = (0.65 - 0.60) / sqrt(0.6*0.4/200) = 1.44. p-value ≈ 0.075. Since 0.075 > 0.05, fail to reject H₀. There isn't enough evidence."
Why it lost points
- Hypotheses H₀ and Hₐ are never written.
- Conditions (random, independent, large sample n·p ≥ 10) are not checked.
- Conclusion is in statistical language but doesn't reference the context (adults exercising).
Strong answer
"H₀: p = 0.60; Hₐ: p > 0.60. Conditions: (1) Random - stated in prompt. (2) Independent - 200 < 10% of all adults. (3) Large counts - np₀ = 120 ≥ 10 and n(1-p₀) = 80 ≥ 10. All satisfied. One-proportion z-test: p̂ = 130/200 = 0.65. z = (0.65 − 0.60) / √(0.6 · 0.4 / 200) = 1.44. p-value = P(Z > 1.44) ≈ 0.0749. Since p-value 0.0749 > α = 0.05, fail to reject H₀. We do not have significant evidence at the 5% level that more than 60% of adults exercise weekly."
Why it scores full marks
- All four steps clearly present: hypotheses, conditions, mechanics, conclusion.
- Conditions explicitly numbered and verified with context.
- Conclusion references both the statistical result AND the context (adults exercising weekly).
Weak vs. strong: confidence interval interpretation
Interpretation prompts are where most students lose points. Readers look for the specific language: 'We are X% confident that the true [parameter in context] is between [L] and [U].'
Prompt
A 95% confidence interval for the mean height of students at a university is (65.2, 67.8) inches. Interpret this interval in context.
Weak answer
"There is a 95% chance that the mean is between 65.2 and 67.8 inches."
Why it lost points
- '95% chance' is wrong - the confidence level refers to the METHOD, not the specific interval.
- No mention of 'true mean' or 'students at this university.'
- Missing the key phrasing readers look for.
Strong answer
"We are 95% confident that the true mean height of all students at this university is between 65.2 and 67.8 inches."
Why it scores full marks
- '95% confident' - the correct frequentist phrasing.
- References 'true mean' (the parameter) explicitly.
- Identifies the population (students at this university) in context.
What you see in GradGPT
This is what your feedback looks like
Every Stats FRQ you write gets scored against the same rubric AP readers use. Strengths, improvements, and notes are highlighted inline.
Your response
p̂ = 0.65. Using a one-proportion z-test, z = 1.44, p-value = 0.075. Since p > 0.05, we fail to reject H₀. There isn't enough evidence.
Inline feedback
Rubric breakdown
You scored higher than 55% of students on this prompt
Hypotheses stated
0/1
H₀ and Hₐ missing
Conditions checked
0/1
Not verified
Mechanics + conclusion
2/2
Calculation + p-value comparison correct
Get this on your own answer.
The 5 FRQ patterns that cover AP Stats
Every AP Stats FRQ fits one of these five patterns. Learn the four-step structure for each and you have the framework for the whole exam.
Hypothesis test (4-step)
Hypotheses → Conditions → Mechanics → Conclusion. Missing any step caps the score.
- Write H₀ and Hₐ in symbols AND in context.
- Verify each condition with a specific check (random, 10%, large counts).
Confidence interval interpretation
One sentence. Specific phrasing required: 'We are X% confident that the true [parameter in context] is between L and U.'
- Always reference 'true' parameter.
- Include context - not just statistical language.
Study design / sampling
Describe a method (SRS, stratified, cluster) or evaluate a design's validity.
- Name the sampling method and how it's executed (random-number table, etc.).
- Explain WHY it reduces bias or protects validity.
Probability / distributions
Binomial, normal, geometric. Name the distribution and parameters explicitly.
- State distribution + parameters: 'X ~ Binomial(n = 50, p = 0.3).'
- Identify the calculator command used (binomcdf, normalcdf) with inputs.
Investigative Task
Open-ended novel application. Readers reward reasoning and connection across multiple concepts.
- Map the familiar technique onto the unfamiliar context.
- Show your reasoning even when the problem is non-standard - label each step.
The mistakes that quietly cost points
These show up every year. Each one is a single habit - fix the habit and you bank points.
Skipping the hypotheses. H₀ and Hₐ are always a separate scoring row on inference FRQs.
Not checking conditions in context. 'Sample is random because…' needs a reason from the prompt.
Interpreting a confidence interval as 'the probability that the true mean is in this interval.' Wrong - it's about the method, not the specific interval.
Writing conclusions in pure statistics language. 'Fail to reject H₀' without context loses the conclusion point.
Forgetting to name the distribution on probability FRQs. 'X ~ Binomial(n, p)' is the notation readers reward.
Rushing the Investigative Task. It's worth more than any short FRQ - spend the full 35 minutes.
Trusted by 10,000+ AP students
What Stats students say after their first graded FRQ
"I tried Barron's, Princeton Review & YouTube. GradGPT-AP is the only thing that actually graded my essays."
Sarah K.
scored a 5
"This is exactly like the real AP exam. Without practicing in GradGPT-AP, I would have run out of time on exam day!"
Marcus T.
3 → 5 on AP Lang in 2 weeks
"Wasted so many hours on Khan Academy. GradGPT-AP showed me exactly what I was doing wrong with my essays. I finally feel confident for the exam."
Priya R.
Your first FRQ scored against the real AP Stats rubric.
Frequently Asked Questions
Six. Five short FRQs (~4 points each) and one Investigative Task (~8 points). You have 90 minutes total - roughly 11 minutes per short FRQ and 35 minutes for the Investigative Task.
Yes. Conditions are almost always a separate scoring row. The three conditions for proportions: random, independent (10% rule), and large counts. For means: random, independent, and Normal/CLT.
Four steps, all present: (1) Hypotheses in symbols AND context. (2) Conditions verified specifically. (3) Test mechanics (statistic, p-value). (4) Conclusion referencing both the statistical result and the context.
Interpretation in pure statistical language. Every conclusion, CI interpretation, or regression inference must be stated in the CONTEXT of the prompt - not just 'we reject' or 'we are confident.'
Very. It's worth roughly twice as much as a short FRQ. Even partial progress scores well if your reasoning is clear and connects familiar concepts to the novel situation.
50/50. Six FRQs carry the same weight as 40 MCQs. FRQs are where most students have the most room to grow, making them the highest-leverage prep.
Write one Stats FRQ. See exactly where you lost points.
Paste your answer and get a rubric breakdown with inline feedback in seconds.
