AP® Biology skills
Confidence intervals and error bars
See what a 95% confidence interval means and when two groups really differ.
About this tool
Turn a mean, a standard deviation and a sample size into a 95% confidence interval, draw it as an error bar, and decide what overlapping or separate error bars tell you.
The formulas, and the rule for reading error bars
SE = s / √n
95% CI ≈ x̄ ± 2 SE
- The 95% confidence interval is the range that very likely contains the true mean of the population the sample came from. On a graph it is drawn as an error bar of ±2 SE around each mean.
- Bars that do not overlap: the difference between the means is likely statistically significant; chance alone would rarely produce samples that far apart.
- Bars that overlap: the data do not show a significant difference. That is not proof that the means are equal: a bigger sample might still find a difference.
- This is the rule of thumb used in this course. A formal test can still find a significant difference when ±2 SE bars overlap slightly; problems here avoid bars that nearly touch.
- Numbers count when they are within one unit of the last decimal place asked for, or within 1%.
Notes for this topic: 95% confidence intervals and error bars