Confidence Interval Calculator (Mean & Proportion)

Choose this endpoint when you need a rigorous statistical confidence interval for either a sample mean (using the t-distribution) or a proportion (using industry-standard Wilson or Wald methods). It is ideal for survey analysis, A/B test result interpretation, quality control, and any scenario where you need to quantify uncertainty around a point estimate. Prefer Wilson intervals for proportions near 0 or 1 or with small sample sizes, as it outperforms Wald in those conditions. This endpoint handles both use cases in one call and returns both Wilson and Wald intervals for proportions, making it more informative than single-method alternatives.

• Missing required 'kind' parameter returns an error
• For 'proportion' kind: missing 'successes' or 'trials' causes an error
• For 'mean' kind: missing or malformed 'values' list causes an error
• Confidence level outside 0.80–0.999 range may be rejected or default to 0.95
• Non-numeric values in the sample list will cause a parse error
• Trials value of zero causes division-by-zero; will return an error or degraded result