Scribe logoScribeNotes & practice
Sign inStart free
  1. Notes
  2. ›AP Statistics
  3. ›3 Inference for Categorical Data: Proportions
  4. ›3.11 Claims from a difference interval

AP Statistics · Unit 3 · Inference for Categorical Data: Proportions

3.11 Claims from a difference interval

Use a confidence interval for a difference of proportions to judge whether two population proportions differ, and interpret zero inside or outside the interval.

  • 3 min read
  • 5 parts
  • 10 questions
  • Draft
Start a Quick 5 on this topic→Jump to the 10 questions
1/5A difference interval can answer a comparison question
  1. 1A difference interval can answer a comparison question
  2. 2An interval containing zero gives no convincing evidence of a difference
  3. 3An interval excluding zero supports a difference with a direction
  4. 4The interval also shows how large the difference might be
  5. 5A complete conclusion states the evidence, the direction, and the context
  6. Check your recall
  7. Practice: 10 questions
1

Part 1 of 5: A difference interval can answer a comparison question

Two sample proportions rarely match exactly. The gap may be real, or it may be sampling variation. A confidence interval for p1−p2p_1-p_2p1​−p2​ helps us judge which.

The key value is zero. A difference of zero means the two population proportions are equal. So we ask whether zero is a plausible value for the difference.

Every value inside the interval is plausible, given the data. Every value outside it is not. We read the interval by checking where zero falls. This gives a direct way to compare two groups.

Key term

plausible value

A plausible value is a number inside a confidence interval that is consistent with the sample data as a possible value of the parameter.

Part 1 done

2

Part 2 of 5: An interval containing zero gives no convincing evidence of a difference

Suppose a 95% interval for p1−p2p_1-p_2p1​−p2​ is (−0.03, 0.11)(-0.03,\,0.11)(−0.03,0.11). Zero lies between the endpoints. So zero is a plausible value for the difference.

Equal proportions are therefore consistent with the data. We do not have convincing evidence that the population proportions differ. The sample gap could be sampling variation. More data might narrow the interval and settle the question.

The interval also holds both positive and negative values. It does not point to a direction either.

Common mistake

Zero inside the interval does not prove the proportions are equal. It means equal is one of several plausible possibilities.

Key term

convincing evidence

Convincing evidence is evidence strong enough that sampling variation alone is an unlikely explanation for the observed result.

Part 2 done

Quick check

One question on the part you just read

Easy

Identify the reference value used to decide whether an interval for p1−p2p_1-p_2p1​−p2​ shows a difference.
Show answer and explanation

Answer: D · Worked explanation

A difference of zero means the two population proportions are equal, so the question is whether zero is plausible. The sample difference sits at the center of the interval and is not the reference. The value 0.5 matters for a single proportion near one-half, not for comparing two proportions.
3

Part 3 of 5: An interval excluding zero supports a difference with a direction

Now suppose the 95% interval is (0.04, 0.18)(0.04,\,0.18)(0.04,0.18). Zero is not between the endpoints. Equal proportions are not plausible, given these data.

So the data provide convincing evidence that the population proportions differ. Every plausible value is positive, so p1>p2p_1>p_2p1​>p2​.

Figure 1. An interval above zero, one containing zero, and one below zero lead to different conclusions about p1−p2p_1-p_2p1​−p2​.

If every plausible value is negative, the evidence points the other way, so p1<p2p_1<p_2p1​<p2​. The order of subtraction fixes the direction. Always read the sign with that order in mind. An interval close to zero still excludes it, but it allows tiny differences.

Part 3 done

4

Part 4 of 5: The interval also shows how large the difference might be

Excluding zero tells us a difference exists. The endpoints also bound how large it might be. An interval of (0.04, 0.18)(0.04,\,0.18)(0.04,0.18) allows differences from 4 to 18 percentage points.

Whether a gap that size matters depends on the context. A lower endpoint close to zero means the true difference could be very small. Report both the evidence and the plausible size when the context asks for it.

Worked example

Writing a conclusion

A 95% confidence interval for pA−pBp_A-p_BpA​−pB​ is (0.053, 0.247)(0.053,\,0.247)(0.053,0.247). Here pAp_ApA​ and pBp_BpB​ are the proportions of all customers who would click design A and design B. Decide whether the interval provides convincing evidence of a difference.
  1. 1
    Locate zero: 0 is below the lower endpoint 0.053, so zero is outside the interval.
  2. 2
    Zero is not plausible, so equal proportions are not consistent with the data.
  3. 3
    Every plausible value is positive, so pA>pBp_A>p_BpA​>pB​.
  4. 4
    Plausible differences run from about 5 to 25 percentage points.

Answer: There is convincing evidence that a larger proportion of all customers would click design A than design B.

Higher confidence levels give wider intervals, so zero may move inside the interval. A conclusion should therefore state the confidence level used.

Part 4 done

Quick check

One question on the part you just read

Easy

Identify the conclusion when a confidence interval for p1−p2p_1-p_2p1​−p2​ contains zero.
Show answer and explanation

Answer: B · Worked explanation

Zero inside the interval means equal proportions are plausible, so there is no convincing evidence of a difference. It does not prove equality, because other plausible values show a difference. A direction cannot be claimed when the interval holds both positive and negative values.
5

Part 5 of 5: A complete conclusion states the evidence, the direction, and the context

Now that you can read an interval, write the conclusion carefully. A complete answer has three parts.

  • Evidence: say whether zero is inside or outside the interval.
  • Conclusion: say whether there is convincing evidence of a difference.
  • Context: name the two populations, and give the direction when zero is outside.

Use the words convincing evidence, not proof. A sample cannot prove a claim about a population.

Exam tip

Match the direction to the order of subtraction. If the interval is for pA−pBp_A-p_BpA​−pB​ and is negative, say that pAp_ApA​ is smaller.

Go deeperWhy intervals and tests usually agree

A confidence interval at level CCC collects the differences that a two-sided test would not reject at significance level 1−C1-C1−C. So a 95% interval that excludes zero matches rejecting equal proportions at the 5% level.

The match is close, not exact. The test combines both samples to estimate one common proportion, while the interval does not.

Note read

Check your recall[12 marks]

6 written questions. Answer each one in your head or on paper, then open the mark scheme.

  1. 1.Identify[1 mark]

    Identify the value that decides whether an interval for p1−p2p_1-p_2p1​−p2​ shows evidence of a difference.

    Show mark schemeHide mark scheme

    Model answer

    Zero: whether zero is inside the interval.

    • 1 mark for identifying zero (whether it lies inside or outside the interval)
  2. 2.Determine[2 marks]

    A 95% confidence interval for p1−p2p_1-p_2p1​−p2​ is (−0.07, 0.05)(-0.07,\,0.05)(−0.07,0.05). Determine whether the interval provides convincing evidence of a difference, and give a reason.

    Show mark schemeHide mark scheme

    Model answer

    The interval does not provide convincing evidence of a difference, because zero is inside the interval and so is a plausible value.

    • 1 mark for concluding that there is no convincing evidence of a difference
    • 1 mark for giving the reason that zero is inside the interval
  3. 3.Interpret[3 marks]

    A 99% confidence interval for pA−pBp_A-p_BpA​−pB​ is (0.02, 0.14)(0.02,\,0.14)(0.02,0.14), where pAp_ApA​ and pBp_BpB​ are the proportions of all patients who recover under treatment A and treatment B. Interpret what the interval tells us about the two proportions.

    Show mark schemeHide mark scheme

    Model answer

    Zero is not in the interval, so there is convincing evidence that the proportions differ. Every plausible value is positive, so a larger proportion of all patients recover under treatment A than under treatment B.

    • 1 mark for noting that zero is outside the interval
    • 1 mark for concluding that there is convincing evidence of a difference
    • 1 mark for stating the direction, pA>pBp_A>p_BpA​>pB​, in context
  4. 4.Explain[2 marks]

    Explain why an interval that contains zero does not show that two population proportions are equal.

    Show mark schemeHide mark scheme

    Model answer

    Zero is only one of many plausible values in the interval. Other values in the interval show a difference, so a difference is still possible.

    • 1 mark for stating that zero is only one of several plausible values
    • 1 mark for explaining that nonzero differences are also plausible, so equality is not shown
  5. 5.Describe[2 marks]

    Describe what an interval such as (−0.20, −0.06)(-0.20,\,-0.06)(−0.20,−0.06) for p1−p2p_1-p_2p1​−p2​ tells us about which population proportion is larger.

    Show mark schemeHide mark scheme

    Model answer

    Zero is outside the interval, so the proportions differ. Every plausible value is negative, so p1<p2p_1<p_2p1​<p2​.

    • 1 mark for stating that zero is outside the interval, giving convincing evidence of a difference
    • 1 mark for stating that all plausible values are negative, so p1p_1p1​ is smaller than p2p_2p2​
  6. 6.Justify[2 marks]

    A student claims that an interval of (0.01, 0.31)(0.01,\,0.31)(0.01,0.31) for pA−pBp_A-p_BpA​−pB​ shows a large difference. Justify why this claim is not supported.

    Show mark schemeHide mark scheme

    Model answer

    The lower endpoint 0.01 means the true difference could be tiny, even though zero is outside. The interval allows both very small and large differences.

    • 1 mark for noting that the interval includes very small differences such as 0.01
    • 1 mark for explaining that excluding zero shows a difference, not that it is large

Practice this topic

10 quick-fire multiple-choice questions, easy to hard. Tap an answer to see if you're right and the worked explanation. (The written questions with mark schemes are under “Check your recall” above.)

0 right0 of 10 answered

Includes the 2 quick checks in the note above.

3.

Easy·Identify[1 mark]

A 95% interval for p1−p2p_1-p_2p1​−p2​ is (0.03, 0.15)(0.03,\,0.15)(0.03,0.15). Identify the correct conclusion.
Show answer and explanation

Answer: C · Worked explanation

Zero is below the lower endpoint, so it is not plausible, and there is convincing evidence of a difference. Every plausible value is positive, so p1>p2p_1>p_2p1​>p2​. The midpoint 0.09 is only the sample difference, not the exact true difference. The sample difference sits at the middle of the interval, and the true value could be anywhere inside.
4.

Medium·Identify[2 marks]

A 95% interval for pA−pBp_A-p_BpA​−pB​ is (−0.18, −0.04)(-0.18,\,-0.04)(−0.18,−0.04). Identify the correct conclusion.
Show answer and explanation

Answer: B · Worked explanation

Zero is above the upper endpoint, so it is not plausible. Every plausible value of pA−pBp_A-p_BpA​−pB​ is negative, so pA<pBp_A<p_BpA​<pB​. The tempting error is to read the negative sign as favoring pAp_ApA​. A negative difference means the first proportion is smaller.
5.

Medium·Identify[2 marks]

A 95% interval for pX−pYp_X-p_YpX​−pY​ is (−0.02, 0.10)(-0.02,\,0.10)(−0.02,0.10). Identify the best statement.
Show answer and explanation

Answer: B · Worked explanation

Zero lies between -0.02 and 0.10, so zero is plausible and there is no convincing evidence of a difference. Having more of the interval above zero does not establish direction. Zero inside never proves equality, since many nonzero differences are also plausible.
6.

Medium·Identify[2 marks]

A 95% interval for pB−pAp_B-p_ApB​−pA​ is (−0.14, −0.02)(-0.14,\,-0.02)(−0.14,−0.02). Identify the correct conclusion about the two proportions.
Show answer and explanation

Answer: B · Worked explanation

Zero is outside the interval, so there is convincing evidence of a difference. Every plausible value of pB−pAp_B-p_ApB​−pA​ is negative, so pBp_BpB​ is smaller and pAp_ApA​ is larger. The order of subtraction is the trap: the first named proportion is pBp_BpB​, not pAp_ApA​.
7.

Medium·Identify[2 marks]

A 95% interval for p1−p2p_1-p_2p1​−p2​ is (0.04, 0.18)(0.04,\,0.18)(0.04,0.18). Identify the statement about the size of the difference that the interval supports.
Show answer and explanation

Answer: A · Worked explanation

The endpoints bound the plausible sizes of the true difference, here 4 to 18 percentage points. The midpoint 0.11 is only the sample difference, not the exact true value. The interval does not say the difference is at least 0.18, and it is not a statement about sample differences.
8.

Hard·Interpret[2 marks]

Using the same data, the 90% interval for p1−p2p_1-p_2p1​−p2​ is (0.012, 0.188)(0.012,\,0.188)(0.012,0.188) and the 99% interval is (−0.038, 0.238)(-0.038,\,0.238)(−0.038,0.238). Interpret the two results.
Show answer and explanation

Answer: C · Worked explanation

A higher confidence level gives a wider interval. The 90% interval excludes zero, but the wider 99% interval includes it, so the conclusion changes with the level. The intervals do not contradict each other, and the wider interval never proves equality.
9.

Hard[2 marks]

A 95% interval for pA−pBp_A-p_BpA​−pB​ is (0.002, 0.030)(0.002,\,0.030)(0.002,0.030). Which statement is most accurate?
Show answer and explanation

Answer: D · Worked explanation

Zero is outside the interval, so there is convincing evidence of a difference, and every plausible value is positive, so pA>pBp_A>p_BpA​>pB​. But the plausible sizes run from 0.2 to 3 percentage points, so the gap may be tiny. Being near zero is not the same as containing zero.
10.

Hard·Determine[2 marks]

Two independent samples give p^1−p^2=0.08\hat{p}_1-\hat{p}_2=0.08p^​1​−p^​2​=0.08, and the margin of error for a 95% interval is 0.10. Determine the correct conclusion.
Show answer and explanation

Answer: D · Worked explanation

The interval is 0.08±0.100.08\pm0.100.08±0.10, which is (−0.02, 0.18)(-0.02,\,0.18)(−0.02,0.18). Zero is inside, so equal proportions are plausible and there is no convincing evidence of a difference. A positive sample difference alone is not enough, and containing zero never proves the proportions are equal.

Keep going

Next topic3.12 Setting up a two-proportion testPrevious topic3.10 Interval for a difference of proportions

More in Unit 3 · Inference for Categorical Data: Proportions

  1. 3.7Carrying out a test for a proportionNotes · 10 questions
  2. 3.8Errors in testsNotes · 10 questions
  3. 3.9Differences in sample proportionsNotes · 10 questions
  4. 3.10Interval for a difference of proportionsNotes · 10 questions
  5. 3.11Claims from a difference intervalNotes · 10 questions
  6. 3.12Setting up a two-proportion testNotes · 10 questions
  7. 3.13Carrying out a two-proportion testNotes · 10 questions
  8. 3.14Setting up a chi-square test for two-way tablesNotes · 10 questions
  9. 3.15Carrying out a chi-square test for two-way tablesNotes · 10 questions

All AP Statistics topics →

Draft: written by Scribe and checked automatically, not yet by a subject reviewer.

Practice questions are original, written by Scribe and checked automatically for correctness.

Topic codes follow the official numbering. Not endorsed by College Board or any exam board.

On this page

0/5 read

  1. 1A difference interval can answer a comparison question
  2. 2An interval containing zero gives no convincing evidence of a difference
  3. 3An interval excluding zero supports a difference with a direction
  4. 4The interval also shows how large the difference might be
  5. 5A complete conclusion states the evidence, the direction, and the context
  6. Check your recall
  7. Practice: 10 questions
All coursesScribeHow it worksPricingPrivacyTerms

Notes and questions are written by Scribe in our own words. Topic codes follow the official course numbering so you can find your place. Not produced or endorsed by the IB, College Board, AQA, OCR, Pearson Edexcel or any exam board.

© 2026 Scribe