top of page

Call or WhatsApp on +919019525923 

Book Free Demo Session

Courses
Grade

AP Statistics Inference Conditions: Units 3 and 4 Guide for 2027

Writer: Aishwarya
Aishwarya
Aug 4
12 min read

Updated: Nov 11

AP Statistics inference (former Units 6 and 7, now Units 3 and 4) requires you to justify the procedure’s conditions using the study design, population information and relevant counts or sample shape. For May 2027, proportions and means are in Units 3 and 4. Name the parameter first, check the conditions that apply and connect each check to the actual problem; a checklist alone does not explain why the method is appropriate.

EduShaale offers personalised 1-on-1 AP Statistics coaching on confidence intervals, tests, condition checks and interpretation. Preparation follows the revised five-unit course and current-format practice questions.


White AP logo with registered trademark symbol on a bright blue background
AP coaching with personalised strategy to score 4s and 5s

Serious About Your AP Scores? Let’s Get You There


EduShaale offers personalised 1-on-1 AP coaching with concept practice and feedback aligned to the current exam.


25–45%

combined MCQ weighting range for current Units 3 and 4

3

procedure-specific checks; not three identical rules for every method

≥10

success/failure threshold for the proportion z procedures below; chi-square checks differ

2.92

the 2025 AP Statistics mean score, out of 5

60.3%

of students scored 3 or higher in 2025

Units 3–4

current locations of proportion and mean inference

Check

connect requested conditions to the problem, rather than asserting they hold

n ≥ 30

current CED sample-size criterion for means; matched pairs use the number of differences

 

Laptop dashboard illustrating AP Statistics inference conditions and data checks

Table of Contents


 


Why Naming a Condition Is Different From Verifying It


In practice, “Random, Independent, Normal” can be a useful reminder, but it is not a substitute for checking the data. A complete explanation identifies the relevant study design and shows how the appropriate numerical or shape condition is satisfied. Use the prompt and scoring criteria to decide what must be written.



The guide keeps its comparisons and worked example while updating the course numbering. Focus on the relevant parameter, the source of the data and the conditions specific to the selected procedure. Proportion counts, mean-distribution checks and matched-pair differences are related ideas with different calculations.


1. Why Conditions Exist -- and Why “SRS” Alone Is Not Enough


Inference procedures have several assumptions, including how data were collected and whether observations can be treated as independent. Normal-approximation checks matter for the z and t procedures discussed here, but they are not the only assumption, and a t statistic follows a t distribution rather than being simply a z score.


The current course includes verifying conditions under Interpret Results, skill 4.E. The ten-point FRQ rubrics award specific components. Check the actual rubric rather than applying the older E/P/I model or assuming every question asks for the same checks.

Ask three questions: how were the data obtained, what supports the required independence assumptions, and is the distribution or count condition appropriate? Random sampling supports population generalisation; random assignment supports treatment comparisons. Those are different roles.


2. A Useful Condition Checklist for z and t Inference


The checklist below is useful for the proportion and mean procedures in this guide. Adapt it to the actual design. Random sampling and random assignment are not interchangeable, and a 10% check is relevant to sampling without replacement.

Condition

What It Checks

How It's Verified

Random

How the data were obtained and what generalisation or treatment comparison is supported

Identify the stated random sample or randomised experiment; do not invent randomisation

Independent

Whether one observation affects the probability of another

Check the design and any between-group independence. When sampling without replacement, verify n ≤ 0.10N using the relevant population.

Normal / Large

Whether the sampling distribution is close enough to normal to trust a z- or t-procedure

Differs by procedure -- this is where proportions and means genuinely diverge (Sections 3 and 4)

The procedure determines the exact checks. Do not assume randomisation automatically proves every kind of independence or that a mean and a proportion use the same distribution criterion.


3. Proportions: Former Unit 6, Current Unit 3


Current Unit 3 includes inference for one proportion and for a difference in proportions. The table shows conventional Large Counts calculations for z procedures. Match the calculation to the procedure; two-sample tests also use a pooled proportion in the test statistic.

Procedure

Large Counts Check Uses

Condition, Written Out

One-proportion confidence interval

Sample proportion, p̂

np̂ ≥ 10 and n(1-p̂) ≥ 10

One-proportion significance test

Hypothesised value, p₀

np₀ ≥ 10 and n(1-p₀) ≥ 10

Two-proportion confidence interval

Each sample's own p̂₁ and p̂₂

n₁p̂₁, n₁(1-p̂₁), n₂p̂₂, n₂(1-p̂₂) all ≥ 10

Two-proportion significance test

Pooled/combined proportion, p̂c

n₁p̂c, n₁(1-p̂c), n₂p̂c, n₂(1-p̂c) all ≥ 10

Checked against current Unit 3. For a two-proportion test, the pooled estimate is p̂c = (x₁+x₂)/(n₁+n₂). The current CED sample scoring notes also accept an observed-count normality check for that test; do not claim a pooled condition check is the only creditworthy approach.


For one-proportion z tests, check counts under p₀. For intervals, use observed successes and failures. A two-proportion test uses the pooled estimate under equal proportions for its test statistic, while current sample scoring allows either the relevant expected or observed counts for checking normality. Follow the method and rubric carefully.


The 10% Condition for Two Samples: Check Both, Separately

When two samples are drawn without replacement, compare each with its own source population. For example, samples of 40 from School A and 35 from School B require suitable population information for each school. Do not combine unrelated populations to make a check appear to pass.


4. Means: Former Unit 7, Current Unit 4


Current Unit 4 covers one mean, independent means and matched-pair differences. Use t procedures when the population standard deviation is unknown. Check the relevant randomisation, sampling and sample-data conditions.

Situation

How the Normal Condition Is Checked

Population is stated to be normally distributed

Condition is automatically satisfied, regardless of sample size

Sample size n ≥ 30

Meets the current CED large-sample criterion for means; still check the other assumptions

Sample size n < 30

The condition depends on the data itself: the sample must be reasonably free of strong skewness and outliers, checked using a dotplot, boxplot, or stemplot

Current CED reference: Unit 4, including 4.2.C and the corresponding test conditions. For matched pairs, the sample size and shape checks concern the differences.

For means, the current CED uses n ≥ 30, not n > 30. If a small sample is used and population normality is not supplied, examine the sample distribution for strong skewness and outliers. A proportion instead needs appropriate success/failure counts.


Matched Pairs: Check the Differences, Not the Two Original Samples

Matched-pairs data -- the same subjects measured twice, such as a class's scores before and after an intervention -- is treated as a single sample of differences, not two separate samples. This has a direct consequence for condition-checking that a meaningful share of students get backwards: the Normal / Large condition is checked on the distribution of the differences, not on the two original sets of scores. Graphing the “before” scores and the “after” scores separately and checking each for skewness answers the wrong question. The one distribution that actually matters is the set of before-minus-after differences for each pair.

For matched-pairs inference about a mean change, define an order of subtraction and form the paired differences. Apply the mean-procedure sample-data checks to those differences. Follow the actual question if it asks you to describe other data as well.


5. Proportions vs Means: The Same Three Conditions, Different Numbers


Proportions and means both require a suitable method and justified assumptions. Naming the parameter first helps you choose the appropriate count or distribution check; the unit number alone is not a method-selection rule.

 

Current Unit 3: Proportions

Current Unit 4: Means

Parameter

p (population proportion)

μ (population mean)

Distribution used

Normal (z)

t; n−1 df for one sample or paired differences; independent two-sample df differ

Normal / Large condition

np ≥ 10 and n(1-p) ≥ 10, using p̂, p₀, or pooled p̂c depending on procedure

Population approximately normal, or n ≥ 30; for n < 30, sample data free of strong skewness and outliers

Formula sheet provided?

Yes

Yes

Common cross-unit error

Using a mean sample-size check instead of the required proportion counts

Applying np ≥ 10 to a mean, where it has no meaning

Name the parameter and design before calculating. That helps distinguish a one-proportion test, an independent two-sample comparison and a paired mean change. Exact credit depends on the question’s criteria.


6. The Four Condition-Checking Mistakes That Cost the Most Points


  1. Naming without verifying: when asked to justify conditions, connect them to the actual design, counts or sample distribution. Not every check needs arithmetic, and a list of words alone may leave the requested explanation incomplete.

  2. Using the wrong proportion: distinguish observed counts for an interval from null-based counts for a one-proportion test. Do not generalise a claim about the most common error without evidence.

  3. Checking the raw data instead of the differences on matched pairs. Graphing or describing the two original sets of scores instead of the paired differences answers a question the procedure never asked.

  4. Combining unrelated populations for a 10% check: each sample drawn without replacement needs its own appropriate population comparison.


7. Same Data, Different Score: A Worked Comparison


The scenario below is original, built to mirror how the College Board's own scoring guidelines treat condition verification, not a reproduction of any specific released question.

Invented practice scenario: a university takes a simple random sample of 150 admitted applicants from a population of 2,000 admitted applicants who received an optional interview. Of these, 102 accept. Test H₀: p = 0.62 against Hₐ: p ≠ 0.62 at α = 0.05, where p is that population’s acceptance proportion. This tests a benchmark, not whether the interview caused a change.

Response A -- Conditions Named, Not Verified

“Conditions: Random ✓, Normal ✓, Independent ✓. z ≈ 1.51, p ≈ 0.130. Fail to reject.”

The numerical result is correct for the revised illustrative scenario, but the explanation does not connect the checks to the sampling design or show the success/failure counts. Its exact credit depends on the requested work and rubric; no official E/P/I classification is assigned here.

Response B -- Conditions Verified With the Actual Numbers

Condition

What Response B Wrote

Random

The problem states a simple random sample of 150 from the 2,000 admitted applicants in the defined population.

Independent

Because sampling is without replacement, check 150 ≤ 0.10(2,000) = 200. Use the actual source population, not applicants nationwide.

Large Counts

Assuming H₀ is true (p = 0.62): np₀ = 150(0.62) = 93 ≥ 10, and n(1-p₀) = 150(0.38) = 57 ≥ 10. Both are satisfied.

Both versions use z ≈ 1.51 and a two-sided p-value ≈ 0.130, so fail to reject H₀. There is insufficient evidence that this population’s acceptance proportion differs from 0.62. The fuller Plan explains the conditions; this invented comparison does not carry a guaranteed exam score.


8. A Step-by-Step Conditions-Checking Framework


Use this practice sequence when a task asks you to justify a proportion or mean inference procedure. Adapt the detail to the question.


  1. Name the parameter first. p or μ? This single decision determines every check that follows.

  2. State Random explicitly, tied to the problem. Not “SRS ✓” but a sentence naming how this specific sample or experiment was actually generated, per the problem text.

  3. Check the independence requirements for the design. If sampling without replacement, compare the sample with its source population; random assignment is a different design feature and does not make every dependence concern disappear.

  4. Choose the procedure-specific check: observed or null-based proportion counts as appropriate, or the population/sample-data criteria for means. For means, the large-sample line is n ≥ 30.

  5. Show relevant evidence: for example, np₀ and n(1−p₀) in a one-proportion test. A study-design statement or a graph-based assessment may also be required; not every condition is numerical.

Practise until you can select and explain the right checks without copying a generic list. Aim for accuracy and relevance before setting a personal time target.


9. Myths About Checking Conditions


Myth: “Listing the three condition names earns the same credit as verifying them.”

Reality: The CED itself flags this as a common and costly shortcut. Verification means connecting each condition to the specific numbers in the problem, not reciting a checklist.


Myth: “The Large Counts condition always uses the sample proportion.”

Reality: One-proportion tests use p₀ for expected counts, while intervals use observed counts. For two-proportion tests, pooled counts are conventional, and the current CED sample rubric also permits observed-count normality checks.


Myth: “If n is at least 30, the Normal condition is automatically fine for any procedure.”

Reality: That threshold applies to means, via the Central Limit Theorem. For proportions, the relevant check is np ≥ 10 and n(1-p) ≥ 10 -- a different calculation with a different threshold.


Myth: “For matched pairs, check both sets of raw scores for Normality.”

Reality: Matched pairs are analysed as a single sample of differences. The Normal / Large condition is checked on those differences, not on the two original datasets.


10. Beyond Proportions and Means: Check the Current Scope


Current Unit 3 retains chi-square tests for homogeneity and independence, which use a different expected-count condition. The goodness-of-fit test and former Unit 9 slope inference were removed from the revised assessed course. Do not carry the z-procedure threshold of 10 into every chi-square check; consult the current CED.

Ready to Start Your AP Journey?


EduShaale offers personalised 1-on-1 AP coaching with concept practice and feedback aligned to the current exam.


Subjects Covered: AP Calculus, AP Physics, AP Chemistry, AP Biology,

AP Economics & more

📞 Book a Free Demo Class: +91 90195 25923 

🌐 www.edushaale.com/advanced-placement  

Free Diagnostic Test: testprep.edushaale.com 

✉️ info@edushaale.com



11. Frequently Asked Questions


Q: What are the three conditions for inference in AP Statistics?

A: For the z and t procedures here, check the randomisation/design, relevant independence or 10% requirement, and the procedure-specific count or sample-data condition. Different procedures do not share one identical universal checklist.

A: For proportion z procedures, check that the relevant successes and failures are at least 10. Use observed counts for intervals and null-based counts for one-proportion tests. For two-proportion tests, current sample guidance also permits observed-count checks.

A: For one proportion, intervals use observed counts and tests use p₀. Two-proportion tests use pooling in the test statistic; the current sample rubric permits relevant expected or observed counts for the normality check.

A: When sampling without replacement, n ≤ 0.10N supports treating observations as approximately independent. Use the sample’s source population and check each independent sample separately where relevant.

A: The current mean criterion uses a normal population, n ≥ 30, or, for n < 30, sample data without strong skewness and outliers. Paired procedures use differences. Proportions need the appropriate counts instead.

A: Matched-pairs data is treated as a single sample made up of the differences between paired measurements, not as two independent samples. Because the inference procedure itself operates on those differences, the Normal / Large condition must be verified on the distribution of the differences, not on the two original sets of scores.

A: If asked to justify conditions, a list of names may leave the required evidence missing. Explain the relevant design or numerical/shape check; exact credit follows that question’s rubric.

A: Condition verification is an assessed skill. Current FRQs use ten-point component scoring; whether a particular point is available depends on the question and its rubric.

A: The same kinds of conditions apply, but two-sample procedures require the relevant checks for both samples rather than once. The 10% condition needs a separate population comparison for each sample, and the Large Counts or Normal condition for two-sample procedures typically involves either four separate counts (proportions) or a check across both samples (means).

A: Check the revised scope. Chi-square homogeneity and independence remain, with their own conditions. Goodness-of-fit and slope inference were removed. Former Unit 6/7 proportion and mean topics are now in Units 3/4.


12. EduShaale -- Expert AP Statistics Coaching


EduShaale offers 1-on-1 AP Statistics coaching on selecting procedures, checking assumptions, calculating and interpreting results for the current exam.


  • Condition-by-Condition Drilling: Every practice free-response question is scored specifically on whether conditions were verified with real numbers, not just named -- the exact distinction this guide is built around.

  • Current inference practice: work on proportions in Unit 3 and means in Unit 4, including matched pairs and independent samples.

  • Rubric-based practice: review current component criteria to identify missing reasoning, calculation or interpretation.

  • Personalised coaching: discuss your current preparation so practice can address conceptual, computational and communication needs.


A useful practice habit is to compare your explanation with the actual condition being checked. Diagnose why a response is incomplete rather than assuming every student’s difficulty has the same cause.


13. References & Resources


Official College Board Resources


EduShaale AP Resources


 

Reviewed 10 November 2026 for May 2027 against College Board’s current Statistics course, exam and Fall 2026 CED. Former Units 6/7 now map to Units 3/4. The 2025 score figures are historical. AP and Advanced Placement are College Board trademarks; College Board did not produce or endorse this guide.

© 2026 EduShaale  |  edushaale.com  |  info@edushaale.com  |  +91 9019525923

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page