In inferential statistics we rarely know the true population parameter. Instead, we collect a sample and use it to estimate the parameter. A single number computed from the sample — such as the sample mean — is called a point estimate. It is our best single guess, but it comes with no built-in measure of uncertainty.
A confidence interval (CI) extends the point estimate into a range of plausible values for the population parameter. Rather than stating “the average lead time is 4.2 days,” we say “we are 95% confident that the true average lead time is between 3.9 and 4.5 days.” The interval communicates both our estimate and the precision of that estimate.
A 95% confidence interval does not mean there is a 95% probability that the true parameter lies inside this specific interval. The true parameter is fixed — it either falls inside or it does not. Instead, the “95%” refers to the procedure: if we were to draw many random samples and build a CI from each one, approximately 95 out of every 100 intervals would contain the true parameter.
Every confidence interval has the form: point estimate ± margin of error. The margin of error captures how far the interval extends on either side of the point estimate. A smaller margin of error means a more precise estimate — but achieving it requires a larger sample size, lower confidence level, or less variability in the data.
Wrong: “There is a 95% chance that the population mean is inside this interval.”
Right: “If we repeated this sampling procedure many times, 95% of the resulting intervals would contain the population mean.” The confidence level describes the long-run success rate of the method, not the probability for any single interval.
When constructing a confidence interval for a population mean, the formula depends on whether the population standard deviation is known or unknown.
When the population standard deviation (σ) is known — rare in practice but common in textbooks — we use the standard normal (Z) distribution. The critical value Z* corresponds to the desired confidence level (e.g., Z* = 1.96 for 95%).
=CONFIDENCE.NORM(alpha, std_dev, n)Z* is the critical value, σ is the population standard deviation, and n is the sample size.
In most real-world situations, we do not know σ and must estimate it with the sample standard deviation s. This additional uncertainty means we use the t-distribution instead of the Z-distribution. The t-distribution is wider (heavier tails), especially for small samples, reflecting the extra estimation error. As the sample size grows, the t-distribution approaches the standard normal.
=CONFIDENCE.T(alpha, std_dev, n)t* is the critical value from the t-distribution with df = n − 1, s is the sample standard deviation, and n is the sample size.
The 95% CI of (3.90, 4.51) days tells operations management that the true average lead time is almost certainly below 4.5 days. If the company's service-level agreement guarantees delivery within 5 days, this interval provides strong evidence that the process is meeting the target on average.
Understanding what makes a confidence interval wider or narrower is critical for designing studies and interpreting results. Three factors control the width of a confidence interval:
Increasing the sample size reduces the standard error (s / n1/2), which shrinks the margin of error. Because n appears under a square root, you must quadruple the sample size to cut the margin of error in half. This diminishing-returns relationship is one of the most important practical insights in statistics.
A higher confidence level (e.g., 99% vs. 95%) requires a larger critical value, which widens the interval. There is always a trade-off between confidence and precision. A 99% CI is wider than a 95% CI built from the same data — you pay for greater confidence with less precision.
More variable data produces wider intervals. If the data values are spread far from the mean, our estimate of the mean is less precise. Reducing measurement variability (through better processes or more consistent sampling) narrows the CI without requiring a larger sample.
The three levers for CI width are sample size, confidence level, and data variability. In practice, analysts most often increase sample size to narrow an interval, since changing the confidence level or reducing inherent variability may not be feasible.
In this chapter we moved from point estimates to interval estimates, learning how confidence intervals quantify the uncertainty inherent in sampling. Here is what you should take away:
Point vs. Interval Estimates: A point estimate is a single number; a confidence interval provides a range of plausible values plus a measure of how confident we are.
Z- and T-Intervals: Use Z when σ is known, t when σ is unknown. The t-interval is wider to account for the extra estimation uncertainty.
Width Factors: Larger n narrows the CI; higher confidence widens it; more variability widens it. Quadrupling n halves the margin of error.
Interpretation: A 95% CI means 95% of similarly constructed intervals would contain the true parameter — it does not mean there is a 95% probability the parameter is in this specific interval.