How should I interpret a 95% confidence interval without saying there is a 95% chance the fixed value is inside it?
Quick answer
A 95% confidence level describes the interval-producing method, not a 95% probability assigned to a fixed parameter after one interval has been calculated. If the method’s assumptions hold, repeating the sampling and interval construction would produce intervals that contain the true parameter about 95% of the time. For a particular calculated interval, say it is a 95% confidence interval, not that the fixed value has a 95% chance of being inside it. [S1]
The question, the details, the next step.
I calculated a 95% confidence interval. Why shouldn’t I say there’s a 95% chance the fixed population value is inside it?
Because the confidence level describes how the method behaves over repeated samples. In the usual interpretation, if you repeatedly took samples and constructed intervals in the same way, about 95% of those intervals would contain the true population parameter, assuming the method’s conditions are satisfied. [S1]
Once you have calculated one interval, its endpoints are fixed, and the population parameter is treated as fixed too. This interpretation does not assign a 95% probability to that parameter being inside this particular interval. You can say, “This is a 95% confidence interval for the population value,” or explain the repeated-sampling meaning. Avoid wording that turns the confidence level into a probability about the already-calculated interval.
Can you show that with a made-up result I could explain to a class?
Suppose, purely as a fictional example, an estimate of a population mean is 13.7 units, and the calculated 95% confidence interval is 12.4 to 15.0 units. A clear explanation is:
“Using this method, the 95% confidence interval for the population mean is 12.4 to 15.0 units. The 95% confidence level means that, over repeated samples, intervals constructed this way would contain the true mean about 95% of the time, if the method’s assumptions hold.” [S1]
The numbers illustrate the wording; they are not real study results. Notice that the explanation makes a statement about the procedure’s long-run coverage. It does not claim that the fixed mean has a 95% chance of being between 12.4 and 15.0.
What are two misleading interpretations I should watch for?
First, avoid: “There is a 95% chance the true mean is in this interval.” That assigns probability to the fixed value after the interval has been calculated.
Second, avoid: “95% of the observations are between the interval endpoints.” A confidence interval for a population mean concerns uncertainty in the mean; it does not say that a specified proportion of individual observations falls within those endpoints. The interval’s meaning depends on the procedure used. The cited explanation addresses a particular method for a population mean, so do not treat it as establishing the assumptions for every interval procedure. [S1]
How can I check my wording before I put it in a report?
Use this short review:
- Identify the parameter the interval concerns, such as a population mean.
- State the interval’s endpoints and units.
- Describe 95% as the long-run coverage of the interval-producing method, conditional on its assumptions.
- Check that you have not described the fixed parameter as having a 95% chance of lying inside this particular interval.
Copyable template: “The [confidence level]% confidence interval for [parameter] is [lower endpoint] to [upper endpoint] [units]. If the method’s assumptions hold, repeated use of this procedure would produce intervals containing the true parameter about [confidence level]% of the time.” For a specific interval method, check its own assumptions; the cited source’s explanation is method-specific and does not establish the conditions or interpretation for every procedure. [S1]
A practical checklist
Name the population parameter, not just the sample estimate.
Give both interval endpoints and the relevant units.
Explain the confidence level as long-run coverage of the procedure, with assumptions stated.
Replace “95% chance the fixed value is inside this interval” with a repeated-sampling interpretation.
Check the specific interval method’s assumptions before applying this wording.
A few more useful answers
Can I say I am 95% confident that the interval contains the true value?
If you use that wording, explain that it refers to the method’s 95% long-run coverage under its assumptions. Avoid presenting it as a probability assigned to the fixed parameter after calculating this particular interval. [S1]
Does a 95% confidence level mean the interval is correct 95% of the time?
It refers to the long-run proportion of intervals from repeated use of the method that contain the true parameter, assuming the method’s conditions hold. It does not certify an individual interval as correct with a 95% probability. [S1]
Does this interpretation apply to every kind of confidence interval?
Do not assume that every interval procedure has the same requirements. The cited explanation treats a particular population-mean setting with stated assumptions; check the method used for your interval. [S1]
Sources
References used for this guide.