Writing an Interval Conclusion in Context

From an Interval to a Statistical Conclusion

A confidence interval is useful only when its numbers are connected to the question that prompted the analysis. Reporting “38.3 to 46.9” leaves a reader to guess what quantity is being estimated, which population is under discussion, and what the numbers measure. A good conclusion makes those details explicit and describes uncertainty without overstating what the interval can establish.

For a one-sample t interval, the target is a population mean, denoted by \(\mu\). The interval is based on the sample mean \(\bar{x}\), the sample standard deviation \(s\), the sample size \(n\), and a t critical value. Its endpoints estimate plausible values of \(\mu\), not the values of individual observations.

Definition: A one-sample t confidence interval for a population mean has the form \(\bar{x} \pm t^* \frac{s}{\sqrt{n}}\), where \(t^*\) is chosen for the stated confidence level with \(df=n-1\). The interval’s two endpoints are estimates for the population mean \(\mu\). A contextual conclusion names the target population, the measured quantity and its units, both endpoints, and the confidence level.

The interval’s interpretation depends on the sampling or model conditions that justify the method. In the examples below, suppose those conditions have been checked and the t interval is appropriate. A conclusion should not imply that the interval itself establishes those conditions.

What a Contextual Conclusion Needs

Before writing, identify the population and the quantity measured. For example, if the data are completion times for a defined group of hikers, the parameter is the mean completion time for that population of hikers. It is not the mean time for a different group, nor does it describe how long every hiker takes.

Next, state the interval with units. Keep units attached to the quantity: minutes, grams, millimetres, or another suitable measure. Round endpoints enough to communicate the result without implying more precision than the data and calculation support. Finally, state the confidence level and use language that describes an estimate rather than a certainty.

Useful form: “We are [confidence level]% confident that the population mean [name the quantity and population] is between [lower endpoint] and [upper endpoint] [units].”

This familiar wording is a conventional summary of the procedure’s uncertainty. Its precise meaning is about repeated sampling: if the same valid interval procedure were used on many samples from the target population, the stated percentage of those intervals would contain the fixed population mean. After one sample has been collected, its endpoints are fixed; the procedure does not assign a probability to the mean being inside this particular interval.

Worked Conclusions

Worked Example: Reporting a Mean Completion Time

Suppose a hypothetical study takes a suitable sample of 16 hikers from a clearly defined population and records their time to complete a trail. The sample mean is 42.6 minutes and the sample standard deviation is 8.0 minutes. Assume the t interval is appropriate. Write a 95% confidence conclusion. For \(df=15\), use \(t^*=2.131\).

The standard error is \(s/\sqrt{n}\), so

SE=\frac{8.0}{\sqrt{16}}=\frac{8.0}{4}=2.0\text{ minutes}.

The margin of error is \(t^*SE\):

ME=2.131(2.0)=4.262\text{ minutes}.

Thus the unrounded endpoints are \(42.6-4.262=38.338\) minutes and \(42.6+4.262=46.862\) minutes. Rounded to one decimal place, the interval is 38.3 to 46.9 minutes.

Conclusion: We are 95% confident that the mean trail-completion time for the defined population of hikers is between 38.3 and 46.9 minutes. The estimate concerns the population’s average time, not a range expected to contain 95% of individual hikers’ times.

Worked Example: Naming the Population and Measurement

In a hypothetical quality-control exercise, a suitable sample of 25 containers is drawn from a production process. Their measured fill depths have a sample mean of 3.84 millimetres and a sample standard deviation of 0.60 millimetres. Assume the t interval is appropriate. Construct and report a 95% interval for the mean fill depth of containers from this process. Use \(t^*=2.064\) for \(df=24\).

First calculate the standard error and margin of error:

SE=\frac{0.60}{\sqrt{25}}=0.12\text{ mm}, \qquad ME=2.064(0.12)=0.24768\text{ mm}.

The endpoints are \(3.84-0.24768=3.59232\) mm and \(3.84+0.24768=4.08768\) mm. Rounding to two decimal places gives 3.59 mm to 4.09 mm.

Conclusion: We are 95% confident that the mean fill depth for containers produced by this process is between 3.59 mm and 4.09 mm. Naming the production process identifies the population of interest; the units make clear what the endpoints measure. This interval does not say that every container’s fill depth lies between those values.

Worked Example: Interpreting an Interval for Study Time

Suppose a hypothetical sample of 10 students from a specified population reports hours spent studying during a particular week. The sample mean is 18.4 hours and the sample standard deviation is 3.2 hours. Assume the conditions for the t interval are justified. Find a 90% interval and write the conclusion. Use \(t^*=1.833\) for \(df=9\).

The standard error and margin of error are

SE=\frac{3.2}{\sqrt{10}}\approx 1.0119\text{ hours}, \qquad ME=1.833(1.0119)\approx 1.8548\text{ hours}.

The endpoints are approximately \(18.4-1.8548=16.5452\) and \(18.4+1.8548=20.2548\) hours. To one decimal place, the interval is 16.5 to 20.3 hours.

Conclusion: We are 90% confident that the mean number of hours studied during the specified week by students in this population is between 16.5 and 20.3 hours. The confidence level is 90%, not 95%; the conclusion reports an estimate of a population average, not a guaranteed range for each student’s study time.

Why the Confidence Wording Is Appropriate

The phrase “95% confident” is not a claim that there is a 95% probability that \(\mu\) falls between the observed endpoints. In the usual frequentist interpretation, \(\mu\) is a fixed population value. It is the interval that would vary from sample to sample. The confidence level describes the long-run performance of the interval-producing method when its assumptions hold.

This interpretation can be expressed using random endpoints. If \(L\) and \(U\) denote the lower and upper endpoints that would result from a random sample, a 95% confidence procedure has coverage

P(L \leq \mu \leq U)=0.95.

The probability statement refers to the procedure before observing the sample. Once the sample is observed, \(L\) and \(U\) are particular numbers. The conclusion then reports those numbers as an interval estimate, while the confidence level describes how the method performs over repeated samples.

The endpoints are also not a prediction interval for individual measurements. An interval for \(\mu\) estimates the population average; individual observations can be substantially lower or higher. Likewise, the sample mean \(\bar{x}\) is the calculated average in the sample, while \(\mu\) is the unknown average in the target population. A contextual conclusion should not switch between these two quantities.

A Short Justification for the Interpretation

A confidence interval is constructed so that the procedure has a specified coverage rate when the model and sampling assumptions hold. For a t interval, the central calculation uses the standardized statistic \((\bar{x}-\mu)/(s/\sqrt{n})\). The chosen critical value gives a central range for this statistic; rearranging that range to isolate \(\mu\) produces the interval endpoints.

-t^* \leq \frac{\bar{x}-\mu}{s/\sqrt{n}} \leq t^* \quad\Longleftrightarrow\quad \bar{x}-t^*\frac{s}{\sqrt{n}} \leq \mu \leq \bar{x}+t^*\frac{s}{\sqrt{n}}.

Because the interval procedure is designed to capture \(\mu\) at the stated long-run rate, it is appropriate to describe the observed interval as a confidence interval for the population mean. This justification does not make the mean random; it explains why the interval procedure has a confidence level.

Common Mistakes and Better Alternatives

Misleading statement	What is wrong	Better approach
“There is a 95% probability that the population mean is in this interval.”	It treats the fixed mean as random after the interval has been observed.	Say that the method captures the fixed mean in 95% of repeated samples, and report the observed endpoints.
“95% of individuals have values between the endpoints.”	The interval estimates a mean; it is not a range for individual values.	Name the population mean as the quantity being estimated.
“The mean is 42.6.”	That reports the sample mean as if it were the known population mean.	Call 42.6 the sample mean and use the interval to express uncertainty about the population mean.
“The mean is between 38.3 and 46.9.”	Without context, the reader cannot tell which mean or what the units are.	Name the population, measured quantity, units, endpoints, and confidence level.

A well-written conclusion is concise, but it is not vague. Include only the population supported by the study design: an interval based on a sample from one defined group does not automatically describe a broader group. Also avoid claiming that an interval proves a practical threshold has been exceeded or that one population differs from another. Those claims require a question and analysis suited to that comparison.

Key takeaway: A contextual one-sample t interval conclusion reports a range for a population mean, not for individual observations. Name the population and quantity, include both endpoints and units, state the confidence level, and interpret that level as a long-run property of the procedure.

Check Your Understanding

For each question, focus on what the interval estimates and how the conclusion should be worded.

A 95% interval for the mean resting pulse rate of a defined population is 68 to 74 beats per minute. Write a contextual conclusion without implying that 95% of individuals’ pulse rates fall in the interval.
Explain the difference between the sample mean \(\bar{x}\) and the population mean \(\mu\) in a confidence-interval conclusion.
Why is “there is a 90% chance that \(\mu\) lies in this observed interval” not the precise frequentist interpretation?
A confidence interval for mean package mass has endpoints of 492 and 508. What additional information should a complete contextual conclusion include?
If a t interval estimates a population mean of weekly exercise time, does it describe the range containing most individuals’ exercise times? Explain.