Margin of error estimates the sampling uncertainty around a proportion at a chosen confidence level.
Margin of error estimates the sampling uncertainty around a proportion at a chosen confidence level.
| Input | What it means |
|---|---|
| Confidence Level | The coverage level used to choose the normal-distribution critical value. |
| Probability | Use the measured or stated value for this item, with the unit or category shown beside the field. |
| Sample size(n) | Number of observations or responses in the sample, rather than the population size. |
| Population size | Total number of eligible members in the population being studied. |
At 95% confidence with p=0.5 and n=400, e≈4.9 percentage points before correction.
Report a sampling-uncertainty band alongside a survey proportion.
A statistic describes the supplied data under stated assumptions. Sampling design, missing observations and the suitability of the model still determine what can reasonably be concluded.
Use the observations required by your study design and state any exclusion rule. Removing values only because they change the answer can bias the analysis.