# Justifying a Claim Based on a Confidence Interval for the Difference Between Two Population Proportions

> AP Statistics · AP 2024-2026 Statistics
> Source: https://www.owlsprep.com/study/ap-statistics-u12-justifying-a-claim-based-on/

This module covers step-by-step validation of statistical claims using two-population proportion confidence intervals, aligned to official AP Statistics free response rubric requirements for full scoring credit.

**Prerequisites:** [Constructing confidence intervals for difference in two population proportions](https://www.owlsprep.com/study/ap-statistics-u12-two-proportion-z-interval/); [Interpreting confidence intervals in context](https://www.owlsprep.com/study/ap-statistics-u11-confidence-interval-interpretation/)

## Learning objectives

- Identify valid statistical claims that can be supported by a two-proportion confidence interval
- Justify whether a difference between two population proportions is statistically significant using interval bounds
- Explain the connection between interval interpretation and contextual claim validation
- Avoid common misinterpretations of two-proportion intervals when evaluating research claims

## Core Rules for Claim Justification with Two-Proportion Intervals

All valid justifications for a claim about $p_1 - p_2$ must explicitly reference the full range of plausible values contained in your calculated confidence interval, rather than only the point estimate.

**Null Difference Value** — The reference value representing no difference between the two population proportions, used to test for evidence of an effect.

*Notation:* 0

1. If 0 is NOT contained in the interval, you have statistically significant evidence at the $\alpha = 1 - CL$ level that the two proportions differ
2. If 0 IS contained in the interval, there is not statistically significant evidence to reject a claim that the two proportions are equal
3. All values inside the interval are plausible for the true difference $p_1 - p_2$, so you cannot rule any of them out

**Worked example:** A 95% confidence interval for $p_{male} - p_{female}$ (proportion who exercise weekly) is [0.03, 0.11]. Justify the claim that males have a higher weekly exercise rate than females.

1. First, note the interval contains only positive values, all of which represent a difference where the male proportion is larger than the female proportion
2. The null difference value 0 is not inside the interval, so no plausible scenario exists where the two proportions are equal
3. We can therefore conclude we have statistically significant evidence at the $\alpha=0.05$ level to support the claim that males have a higher weekly exercise rate

**Check your understanding**

Test your basic understanding

1. A 90% confidence interval for $p_{control} - p_{treatment}$ is [-0.02, 0.07]. Is there evidence the treatment changes the outcome proportion?

   *Why:* All values from -0.02 to 0.07 are plausible, including no difference at 0, so you cannot support a claim that the treatment has an effect.

## Evaluating Directional vs Non-Directional Claims

Many AP exam prompts ask you to justify a one-sided (directional) claim, rather than just a general 'they differ' claim. You must confirm the entire interval aligns with the direction of the claim to support it.

**Exam command terms**

AP exam command terms for this topic have strict rubric requirements:

- **Justify the claim** — You must explicitly reference the interval's contents, not just state a conclusion *(Because all values in the 95% interval are negative, we support the claim $p_1 < p_2$)*

- **Explain if the interval provides evidence** — You must name the null value 0 and state whether it falls inside the interval *(Since 0 is contained in the interval, there is no evidence the proportions differ)*

**Worked example:** A 99% confidence interval for $p_{senior} - p_{junior}$ (proportion who own a car) is [-0.14, 0.02]. A student claims seniors have a lower car ownership rate than juniors. Justify if this claim is supported.

1. The interval ranges from -0.14 (senior proportion 14% lower) to +0.02 (senior proportion 2% higher)
2. The interval contains 0 and multiple positive values, meaning it is plausible that seniors have equal or even higher car ownership rates than juniors
3. We do not have sufficient evidence to support the student's directional claim, as not all plausible values align with the claim

## Full AP Rubric-Aligned Response Structure

To earn full credit on AP Free Response questions for this topic, your response must hit three distinct scoring points that graders are explicitly instructed to look for.

- 1. Restate the given confidence interval in the context of the problem
- 2. Explicitly state whether the null difference value 0 is inside or outside the interval
- 3. Connect that observation directly to the original claim, using contextual language not just statistical jargon

> **Full Credit Mnemonic**
>
> Use the 3-step mnemonic 'Context → Zero → Conclusion' to never miss a scoring point

**Worked example:** Prompt: A school calculates a 95% confidence interval for $p_{remote} - p_{inperson}$ (proportion of students who pass the exam) as [-0.08, 0.01]. Justify the claim that remote learning reduces pass rates.

1. Context step: This interval estimates the difference in pass rates between remote and in-person students, ranging from an 8% lower pass rate for remote students to a 1% higher pass rate for remote students
2. Zero step: The value 0, representing no difference in pass rates, is contained inside this interval
3. Conclusion step: Since it is plausible there is no difference at all, we do not have evidence to support the claim that remote learning reduces pass rates

## Handling Claims About Exact Magnitudes of Difference

Some advanced prompts ask you to evaluate a claim that the difference between two proportions is exactly equal to a specific non-zero value, for example 'the difference in approval rates is 10%'.

The same rule applies: if the exact claimed value is inside your confidence interval, the claim is plausible; if it is outside the interval, you have evidence against that specific claim.

**Worked example:** A poll calculates a 95% confidence interval for $p_{democrat} - p_{republican}$ (voter approval) as [0.04, 0.12]. A politician claims the true difference in approval is 15%. Is this claim supported?

1. The interval of plausible differences runs from 4 percentage points to 12 percentage points
2. The claimed value of 15% (0.15) falls outside the upper bound of the interval
3. We have statistically significant evidence at the $\alpha=0.05$ level to reject the politician's claim, as 15% is not a plausible value for the true difference

## Common pitfalls

- **Wrong:** Saying 'the interval proves the two proportions are equal' when 0 is inside the interval
  - Why it fails: A confidence interval never proves equality, it only fails to find evidence of a difference
  - Correct: State that there is not sufficient evidence to conclude the proportions differ
- **Wrong:** Only referencing the point estimate to justify a claim, ignoring the full interval range
  - Why it fails: The point estimate is just one plausible value, you must account for all values in the interval
  - Correct: Explicitly reference the full lower and upper bounds of the interval in your justification
- **Wrong:** Using a 95% interval to justify a claim of 100% certainty that the difference is positive
  - Why it fails: Confidence intervals only have 95% confidence, not 100%, so you cannot state absolute certainty
  - Correct: Frame your conclusion as statistically significant evidence supporting the claim, not absolute proof
- **Wrong:** Switching the order of the two proportions mid-justification, flipping the sign of the difference
  - Why it fails: Flipping $p_1 - p_2$ to $p_2 - p_1$ reverses the direction of your conclusion and loses all credit
  - Correct: Restate the definition of $p_1$ and $p_2$ at the start of your justification to avoid mix-ups
- **Wrong:** Claiming you can generalize the conclusion to populations outside your sampling frame
  - Why it fails: Your interval only applies to the two populations you sampled from, not unrelated groups
  - Correct: Restrict your conclusion explicitly to the two target populations defined in the study

## Cheatsheet

| Scenario | Is 0 inside interval? | Claim Justification Outcome |
| --- | --- | --- |
| Non-directional claim: $p_1$ differs from $p_2$ | No | Significant evidence the proportions differ |
| Non-directional claim: $p_1$ differs from $p_2$ | Yes | No statistically significant evidence of a difference |
| Directional claim: $p_1 > p_2$ | Interval contains any negative or zero values | Cannot support the directional claim |
| Directional claim: $p_1 > p_2$ | All interval values are positive | Significant evidence to support the directional claim |
| Exact magnitude claim: $p_1 - p_2 = k$ | k is inside the interval | The claimed value k is plausible |
| Exact magnitude claim: $p_1 - p_2 = k$ | k is outside the interval | Evidence against the claimed value k |

## What's next

Mastering this justification skill is one of the highest-yield moves for your AP Stats exam, as it appears in nearly every free response set focused on two-proportion inference. You will now be able to earn full scoring points for any prompt that asks you to connect a calculated interval to a real-world research claim, rather than losing partial credit for skipping required rubric steps. This skill also directly transfers to justifying claims for the difference between two population means, which you will cover next in your inference unit. Practice applying this 3-step structure to every two-proportion interval problem you encounter to build automaticity for exam day.

---

From [OwlsPrep](https://www.owlsprep.com) — free study guides for A-Level, IB, AP and IGCSE, written against the official syllabus. Canonical page: https://www.owlsprep.com/study/ap-statistics-u12-justifying-a-claim-based-on/