Large Counts Condition

The Large Counts Condition says the sampling distribution of a sample proportion p̂ is approximately normal when the expected number of successes and failures are both at least 10 (np ≥ 10 and n(1−p) ≥ 10), which justifies using z-procedures for proportions.

What is the Large Counts Condition?

The Large Counts Condition is the normality check for proportions. Before you can use a normal (z) distribution to do inference about a population proportion, you need the sampling distribution of p̂ to actually look normal. Per the CED (5.5.B), that happens when the sample is big enough that you expect at least 10 successes AND at least 10 failures, written as np ≥ 10 and n(1−p) ≥ 10. Why 10? With too few successes or failures, the distribution of p̂ gets skewed and lumpy, and a normal curve would give you bad probabilities.

The tricky part is that which proportion you plug in changes depending on the procedure. In Unit 5 and for confidence intervals, you use the sample proportion p̂ (check np̂ ≥ 10 and n(1−p̂) ≥ 10). For a one-sample z-test, you assume H₀ is true, so you use the hypothesized value p₀ (check np₀ ≥ 10 and n(1−p₀) ≥ 10). For a two-sample test of p₁ − p₂, you use the combined (pooled) proportion p̂c for all four counts. For a two-sample interval, you check all four counts with each sample's own p̂: n₁p̂₁, n₁(1−p̂₁), n₂p̂₂, and n₂(1−p̂₂).

Why the Large Counts Condition matters in AP Statistics

This condition first shows up in Topic 5.5 (learning objective 5.5.B), where you decide whether the sampling distribution of p̂ can be described as approximately normal. Then it becomes a required step in every proportion inference procedure in Unit 6: verifying conditions for a one-proportion z-test (6.4.C), a two-proportion z-interval (6.8.B), and a two-proportion z-test (6.10.C). The whole logic of z-intervals and z-tests rests on the normal approximation. If Large Counts fails, the z-procedure isn't valid, and the AP exam expects you to say so. On FRQs, 'check conditions' is graded as its own component, so skipping or fumbling this check costs real points even when your calculations are perfect.

How the Large Counts Condition connects across the course

10% Condition (Units 5-6)

These two conditions travel together, but they check different things. The 10% Condition (n ≤ 10% of the population) protects independence when sampling without replacement, so the standard deviation formula works. Large Counts protects the shape, so the normal curve works. You need both, and they are not interchangeable.

Sampling Distributions for Sample Proportions (Unit 5)

Topic 5.5 is where Large Counts is born. The sampling distribution of p̂ is built from a binomial count divided by n, and binomial distributions only look bell-shaped when successes and failures both pile up to at least 10. Unit 6 just keeps reusing this fact.

1-Prop Z-Test (Unit 6)

Here's the detail that separates 4s from 5s. In a hypothesis test you assume H₀ is true, so you check Large Counts with p₀, the hypothesized proportion, not the p̂ you observed. Writing np̂ ≥ 10 on a test setup is a classic rubric-losing move.

Confidence Interval for a Difference of Two Proportions (Unit 6)

With two samples, Large Counts becomes four checks instead of two. For an interval you verify n₁p̂₁, n₁(1−p̂₁), n₂p̂₂, and n₂(1−p̂₂) are all at least 10. For the matching two-sample test, you swap in the pooled proportion p̂c for all four.

Is the Large Counts Condition on the AP Statistics exam?

On multiple choice, expect stems like 'Which condition is necessary for the sampling distribution of p̂ to be approximately normal?' or scenarios where one condition fails and you have to spot which one. On free-response, Large Counts shows up inside every 'construct and interpret a confidence interval' or 'perform a test' question about proportions. A complete answer names the condition, shows the actual arithmetic (like np₀ = 200(0.4) = 80 ≥ 10 and n(1−p₀) = 200(0.6) = 120 ≥ 10), and states the conclusion that the sampling distribution is approximately normal. Just writing 'Large Counts: ✓' without numbers typically doesn't earn the conditions component. Also be ready to use the right proportion for the right procedure (p̂ for intervals, p₀ for one-prop tests, p̂c for two-prop tests).

The Large Counts Condition vs 10% Condition

Both are conditions you check before proportion inference, but they answer different questions. The 10% Condition (sample size at most 10% of the population) keeps observations approximately independent when you sample without replacement, which makes the standard deviation formula valid. The Large Counts Condition (at least 10 successes and 10 failures) makes the shape of the sampling distribution approximately normal, which makes z-scores and z* critical values valid. A quick memory hook: 10% is about independence, 10 counts is about normality.

Key things to remember about the Large Counts Condition

• The Large Counts Condition requires at least 10 expected successes and at least 10 expected failures (np ≥ 10 and n(1−p) ≥ 10) for the sampling distribution of p̂ to be approximately normal.

• Which proportion you plug in depends on the procedure: use p̂ for a one-sample confidence interval, p₀ for a one-sample z-test, each sample's p̂ for a two-sample interval, and the pooled proportion p̂c for a two-sample test.

• For two-sample procedures, Large Counts means checking four counts, not two, and all four must be at least 10.

• Large Counts checks shape (normality), while the 10% Condition checks independence; you need both, and they are not the same thing.

• On FRQs, show the actual numbers when verifying the condition; a bare checkmark next to 'Large Counts' usually won't earn credit.

• If the Large Counts Condition fails, the normal approximation isn't valid and you cannot justify using a z-interval or z-test for the proportion.

Frequently asked questions about the Large Counts Condition

What is the Large Counts Condition in AP Stats?

It's the check that a sample has at least 10 expected successes and 10 expected failures (np ≥ 10 and n(1−p) ≥ 10). When it's met, the sampling distribution of the sample proportion p̂ is approximately normal, so z-procedures for proportions are valid.

Do I use p̂ or p₀ when checking the Large Counts Condition?

It depends on the procedure. For a confidence interval, use the sample proportion p̂. For a one-proportion z-test, assume H₀ is true and use the hypothesized value p₀, checking np₀ ≥ 10 and n(1−p₀) ≥ 10. For a two-proportion z-test, use the pooled proportion p̂c.

Is the Large Counts Condition the same as the 10% Condition?

No. The 10% Condition (n ≤ 10% of the population) checks independence when sampling without replacement. Large Counts (at least 10 successes and 10 failures) checks that the sampling distribution's shape is approximately normal. Proportion inference requires both.

Why does the Large Counts Condition use 10?

Ten is the conventional cutoff the CED uses to guarantee the binomial-based distribution of p̂ isn't too skewed for a normal approximation. With fewer than 10 expected successes or failures, the distribution is noticeably skewed and normal probabilities would be inaccurate.

Does the Large Counts Condition apply to means too?

No, it's specific to proportions (categorical data). For means of quantitative data, normality is justified differently, usually through a normal population, a sample size of about 30 or more via the Central Limit Theorem, or a graph showing no strong skew or outliers.