What Is a Confidence Interval?
The short answer is that a confidence interval wraps a guess in a margin of error. You take a sample, get a number like an average, then add and subtract a margin to build a range. The confidence level tells you how often the recipe works over many samples.
What does a 95% confidence interval mean?
It does not mean there is a 95% chance the true value sits in your specific interval. That is the most common mix-up. The 95% describes the long run: if you repeated your sampling many times, about 95% of the intervals you build would contain the true value.
Once you have one interval, it either contains the true number or it does not. You just do not know which. Statistician Jerzy Neyman built the idea this way in the 1930s, and it still trips people up.
What is the confidence interval formula?
For a sample average, the standard formula is:
CI = x̄ ± z × (s / √n)
- x̄ is your sample average
- z is the critical value, about 1.96 for 95% confidence
- s is the sample standard deviation
- n is the sample size
The chunk after the ± sign is the margin of error. Bigger samples shrink it. A higher confidence level widens it.
| Confidence level | z value |
|---|---|
| 90% | 1.645 |
| 95% | 1.96 |
| 99% | 2.576 |
How do you work out a confidence interval?
Say you weigh 25 apples and get a mean of 180 grams with a standard deviation of 10 grams. Plug in the numbers:
Margin = 1.96 × (10 / √25) = 1.96 × 2 = 3.92
So the 95% confidence interval is roughly 176.1 to 183.9 grams. You can report it as 180 ± 3.9 grams.
Why do confidence intervals matter?
A single average hides its own uncertainty. The interval shows how much wiggle room is honest. A poll saying a candidate leads 51% to 47% looks like a win until you see the margin of error is 4 points, which makes it a coin flip. That is why newspapers now publish intervals next to survey results.
What makes a confidence interval narrower?
- More data: n in the denominator pulls the margin down
- Less variation: smaller s means a tighter range
- A lower confidence level: trade certainty for precision
Do the math: open the confidence interval