AP Statistics · Unit 4 · Inference for Quantitative Data: Means
4.8 Claims about a difference of means
Draw conclusions about a difference of two means from an interval, name the plausible values, and recognize the limits of those conclusions.
1/5Every value inside the interval is a plausible difference
- Every value inside the interval is a plausible difference
- Whether the interval contains zero determines the claim
- An interval gives the same conclusion as a two-sided test
- The endpoints show how large the difference could be
- Conclusions are limited by the design, the conditions and the method
- Check your recall
- Practice: 10 questions
Part 1 of 5: Every value inside the interval is a plausible difference
Imagine a company comparing two battery designs. A study gives a 95% interval of 1.2 to 4.8 hours for . This is the true mean life of design A minus design B.
Values between 1.2 and 4.8 are each a plausible value for that difference. The data are consistent with any of them. Values outside the interval, such as 0 or 6, are not supported by the data at this confidence level.
The sample difference sits at the center of the interval. The true difference could be anywhere inside it.
Key term
plausible value
A plausible value is a value of the parameter that is consistent with the sample data, meaning it falls inside the confidence interval.
Part 2 of 5: Whether the interval contains zero determines the claim
The next question is whether the groups differ at all. If the true means were equal, would be 0, which is the natural reference point. That claim is the hypothesized difference, and you compare it with the interval.
Reading an interval for the difference
- Interval
- All values positive
- Position of 0
- Below the interval
- Claim
- Convincing evidence that
- Interval
- All values negative
- Position of 0
- Above the interval
- Claim
- Convincing evidence that
- Interval
- Both signs present
- Position of 0
- Inside the interval
- Claim
- No convincing evidence of a difference
The sign gives the direction. Positive values mean group 1 has the larger mean, and negative values mean group 2 does.
Key term
hypothesized difference
The hypothesized difference is the value of stated in the null hypothesis, usually 0.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: A · Worked explanation
Part 3 of 5: An interval gives the same conclusion as a two-sided test
This rule matches the logic of a hypothesis test. A confidence interval and a two-sided test use the same data. A interval matches a test at significance level , so a 95% interval matches .
If the hypothesized difference lies outside the interval, reject . If it lies inside, fail to reject . An interval also has an advantage over a test, because it shows how large the difference could be.
Worked example
Deciding from a given interval
- The hypothesized difference is 0.
- Zero lies between -0.8 and 3.6, so it is a plausible value.
- This matches failing to reject at .
Answer: There is no convincing evidence that the true mean study times differ. Plausible differences run from 0.8 hours lower to 3.6 hours higher at school A.
In the example, the confidence level is 95%. Intervals built this way would capture the true difference in about 95% of repeated samples.
Key term
confidence level
The confidence level is the long-run percentage of intervals, built the same way from repeated random samples, that capture the true parameter.
Part 4 of 5: The endpoints show how large the difference could be
Excluding 0 is not the end of the story. The size of the difference also matters.
Suppose a 95% interval for the difference in mean scores is 2.1 to 9.5 points. Every plausible value is positive, so group 1 has the higher mean. Even so, the true gap could be small or large.
Whether a difference is practically significant depends on the context, not on the interval alone. If only gaps above 5 points matter, this interval does not settle the issue. It includes values below 5.
Key term
practically significant
A difference is practically significant when it is large enough to matter in the real-world context of the study.

Quick check
One question on the part you just read
Easy
Show answer and explanation
Answer: A · Worked explanation
Part 5 of 5: Conclusions are limited by the design, the conditions and the method
Even a well-built interval supports only the claims that the study design allows. Two design features matter most, and the conditions matter too.
- Random sampling allows generalizing to the populations sampled.
- Using random assignment allows cause-and-effect conclusions about the treatments.
- Without random assignment, a confounding variable may explain the gap.
The conclusion also depends on the conditions. If a sample is not random, or small samples are strongly skewed, the stated confidence level may be wrong. Finally, the interval describes population means. It says nothing about any single individual, and individual values vary much more than means do.
Key terms · tap one for its definition
random assignment
Random assignment is the use of chance to place subjects into treatment groups, which allows cause-and-effect conclusions about the treatments.
confounding variable
A confounding variable is a variable related to both group and response whose effect cannot be separated from the group effect.
Check your recall[12 marks]
5 written questions. Answer each one in your head or on paper, then open the mark scheme.
- 1.Determine[2 marks]
A 95% interval for is . Determine what the interval suggests about whether the population means differ.
Show mark schemeHide mark scheme
Model answer
Zero lies inside the interval, so it is a plausible value. There is no convincing evidence that the population means differ.
- 1 mark for noting that 0 is inside the interval
- 1 mark for concluding there is no convincing evidence of a difference
- 2.Interpret[3 marks]
Interpret the 95% interval days for , the true mean recovery time for a drug group minus a placebo group.
Show mark schemeHide mark scheme
Model answer
We are 95% confident that the true mean recovery time with the drug is lower than with the placebo by between 1.5 and 6.5 days.
- 1 mark for stating 95% confidence
- 1 mark for giving the direction as lower for the drug group
- 1 mark for giving the size as 1.5 to 6.5 days
- 3.Explain[2 marks]
Explain why a 95% interval that excludes 0 leads to the same conclusion as a two-sided test at .
Show mark schemeHide mark scheme
Model answer
The hypothesized difference of 0 is outside the interval, so it is not plausible. A 95% interval matches , so is rejected.
- 1 mark for stating that 0 is not a plausible value
- 1 mark for linking a 95% interval to and rejecting
- 4.Describe[3 marks]
Describe three limitations on the conclusions drawn from an interval for based on observational data.
Show mark schemeHide mark scheme
Model answer
Without random assignment, a confounding variable may explain the difference, so cause and effect cannot be concluded. The interval is valid only if the conditions hold. The interval describes population means, not individuals.
- 1 mark for no cause-and-effect conclusion without random assignment, because of possible confounding
- 1 mark for the conditions must hold for the interval to be valid
- 1 mark for the interval describes means, not individuals
- 5.Justify[2 marks]
A 95% interval for the difference in mean scores is points. Justify why this does not prove a large difference.
Show mark schemeHide mark scheme
Model answer
The interval excludes 0, but it includes values as small as 0.3 points. The true difference could be small, so a large difference is not established.
- 1 mark for noting that the interval includes small values such as 0.3
- 1 mark for concluding the true difference could be small, so a large difference is not shown

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.
EasyIdentify[1 mark]
Show answer and explanation
Answer: A · Worked explanation
MediumInterpret[1 mark]
Show answer and explanation
Answer: A · Worked explanation
MediumDetermine[1 mark]
Show answer and explanation
Answer: A · Worked explanation
MediumIdentify[1 mark]
Show answer and explanation
Answer: A · Worked explanation
MediumDetermine[1 mark]
Show answer and explanation
Answer: A · Worked explanation
HardIdentify[2 marks]
Show answer and explanation
Answer: A · Worked explanation
HardIdentify[2 marks]
Show answer and explanation
Answer: A · Worked explanation
HardIdentify[2 marks]
Show answer and explanation
Answer: A · Worked explanation