Put the Groups and the Order Into Words
In “Worked Example: Interval for Two Group Means,” you constructed a two-sample t confidence interval for \(\mu_1-\mu_2\). Now the important task is to explain what that interval estimates. A correct interpretation names the two populations, states which mean is subtracted from which, and includes the variable and its units.
The order matters because \(\mu_1-\mu_2\) means the true population mean for group 1 minus the true population mean for group 2. It does not mean the reverse difference, and it is not an interval for either group’s mean on its own.
A Contextual Sentence That Covers the Essentials
A reliable interpretation has a simple structure. First state the confidence level. Then name the true mean in the first population minus the true mean in the second population. Finally, give the interval endpoints and the units. Keep “minus” visible in the sentence so the reader can tell which group is first.
For example, if the interval is for average water use in one type of household minus average water use in another type, do not shorten the interpretation to “the difference is between the endpoints.” Say which households are first and which are second. That makes the estimated quantity unambiguous.
Both endpoints use the same units as the original quantitative variable. If the variable is measured in minutes, the endpoints describe a difference in minutes; if it is measured in kilograms, the endpoints describe a difference in kilograms. If the variable is recorded as a percentage, the interval is typically expressed in percentage points, not as a relative percent difference; if the variable is recorded as a proportion, the interval is in proportion units.
The confidence level describes the long-run success rate of the interval-producing method. If we repeatedly took appropriate samples and constructed intervals in the same way, about the stated percentage of those intervals would capture the true difference \(\mu_1-\mu_2\). The population means are fixed; the intervals vary from sample to sample.
So, for a particular interval, avoid saying there is a 95% probability that the fixed difference is inside it. The AP-standard interpretation is “We are 95% confident that...” followed by the parameter in context and the interval’s endpoints.
Worked Examples
Worked Example: Household Water Use
An invented environmental project compares weekly household water use in two types of homes. Population 1 is households with a rain barrel; population 2 is households without a rain barrel. The quantitative variable is weekly water use, measured in hundreds of liters. After the two-sample t interval was calculated, the reported 95% confidence interval for \(\mu_1-\mu_2\) was \((-3.2,\ 1.1)\) hundreds of liters.
Here, \(\mu_1\) is the true mean weekly water use for all households with a rain barrel, and \(\mu_2\) is the true mean weekly water use for all households without one. The subtraction order is rain-barrel households minus households without a rain barrel. The interval’s endpoints are already expressed in the variable’s units.
A contextual interpretation is: We are 95% confident that the true mean weekly water use for households with a rain barrel minus the true mean weekly water use for households without a rain barrel is between \(-3.2\) and \(1.1\) hundreds of liters.
This sentence identifies the confidence level, both populations, the variable, the subtraction order, the endpoints, and the units. It describes plausible values for the population mean difference—not the range of water use among individual households and not an interval for either population mean alone.
Worked Example: Battery Life for Two Device Models
An invented consumer-technology project compares battery life for two device models under a specified testing routine. Population 1 is all devices of model A under that routine; population 2 is all devices of model B under the same routine. Battery life is measured in hours. A calculated 90% confidence interval for \(\mu_1-\mu_2\) is \((0.4,\ 1.7)\) hours.
The parameter is the true mean battery life for model A minus the true mean battery life for model B. A suitable interpretation is: We are 90% confident that the true mean battery life for model A minus the true mean battery life for model B, under the specified testing routine, is between 0.4 and 1.7 hours.
Notice that the sentence retains model A first and model B second. Saying “model B minus model A” would interpret a different parameter. The units are hours because the original variable is battery life in hours. The 90% confidence level describes the long-run performance of the method used to produce intervals, not a probability assigned to this particular fixed mean difference.
Worked Example: Recovery Time After Two Care Plans
An invented health project compares recovery time, in days, for people following two care plans. Population 1 is people using plan Cedar; population 2 is people using plan Maple. The reported 95% confidence interval for \(\mu_1-\mu_2\) is \((-2.6,\ 0.8)\) days.
Define the target in words before interpreting the endpoints: it is the true mean recovery time for all people using plan Cedar minus the true mean recovery time for all people using plan Maple. Therefore, the interpretation is: We are 95% confident that the true mean recovery time for people using plan Cedar minus the true mean recovery time for people using plan Maple is between \(-2.6\) and \(0.8\) days.
The endpoints describe a difference in mean recovery times, in days. They do not describe the recovery time of an individual person. The sentence also does not claim that the care plan caused a difference; conclusions about cause depend on how the study was designed. The interpretation’s job is to describe the parameter estimated by the interval.
Check the Sentence Before You Submit It
Use this short review to catch wording problems before you report a confidence interval. It follows the same parameter and interval conventions used in “The Form of a Two-Sample t Interval” and “Interpreting a Paired t Interval,” while keeping the focus on two independent population means.
Translate \(\mu_1-\mu_2\) as the true mean for group 1 minus the true mean for group 2.
Use the variable from the study, such as weekly water use or battery life, rather than saying only “the difference.”
Keep the endpoints in lower-to-upper order and attach the variable’s units to the difference.
Say “We are [confidence level] confident that...” and interpret the interval as plausible values for the fixed population mean difference.
A useful final check is to imagine swapping the group labels. If that swap would change the parameter, the sentence must make clear which group comes first. In particular, do not rely on labels such as “group 1” and “group 2” unless the groups have already been clearly identified in context.
Common Mistakes and AP Exam Tips
- Reversing the subtraction order: For an interval estimating \(\mu_1-\mu_2\), describe group 1 first and group 2 second. Reversing the words changes the quantity being interpreted.
- Interpreting one mean instead of a difference: The interval is not a range for the mean of group 1 by itself. It estimates the mean for group 1 minus the mean for group 2.
- Leaving out the variable or units: “The difference is between 0.4 and 1.7” is incomplete without saying what is being measured and the units of the difference.
- Assigning probability to a fixed parameter: Do not say “There is a 95% probability that the true difference is in this interval.” Explain that we are 95% confident, reflecting the long-run success of the interval method.
- Making a causal claim from the interval alone: An interval estimates a population difference. Whether a study supports a cause-and-effect conclusion depends on its design, not just the interval’s endpoints.
- Reporting only the center: The sample difference in means is not the full confidence interval. Your interpretation should describe the range of plausible values shown by both endpoints.
For full-credit communication, match the order in your sentence to the order in the parameter, name both populations and the quantitative variable, report the endpoints with units, and use the stated confidence level accurately. “We are 95% confident that the true mean for population 1 minus the true mean for population 2 is between...” is a strong start, but the context must identify what those populations and means represent.
Check Your Understanding
For each question, focus on defining the population mean difference and expressing it clearly in context.
- An interval estimates mean commute time for people who bike to work minus mean commute time for people who take a bus. Which group must be named first in the interpretation?
- A 90% interval for mean daily screen time, group A minus group B, is \((-0.5,\ 1.8)\) hours. Write a complete contextual interpretation using those endpoints.
- Why is “There is a 95% probability that the true difference is in this interval” not the preferred interpretation of a 95% confidence interval?
- If the variable is the amount of fertilizer used, measured in kilograms, what units should the endpoints of an interval for \(\mu_1-\mu_2\) have?
- What population quantity does an interval for \(\mu_1-\mu_2\) estimate, and how is it different from either population mean alone?