Sample Size Planning Tool: Precision-Based Design

Plan your sample size based on desired precision :


Enter study parameters

50%
to participants

What precision would different sample sizes get you?

Estimating precision from sample size

With a sample size of and expected success rate around 50%, we can measure success rate with a margin of error of ±--% (confidence interval half-width)

If comparing two groups...

For comparing two conditions with 16 participants:

  • We can detect differences of ±--% or greater with statistical significance (95% CI excluding zero)
  • We can be 95% confident about which condition is better when the true difference is ±--% or greater
  • We can be 80% confident about which condition is better when the true difference is ±--% or greater

Understanding the Analysis

This tool uses a Bayesian approach with beta-binomial distribution to calculate:

  • Credible intervals (Bayesian version of confidence intervals) for more intuitive probability statements
  • Minimum detectable differences between treatments, both for statistical significance and directional confidence
  • How precision improves with increasing sample size

All calculations use a uniform prior, which makes minimal assumptions about the outcomes before collecting data.