MDE mínimo detectável
- Created by
- Renato Passos, Eng. de Software
- Reviewed by
- Renato Passos, Eng. de Software
Last updated: Apr 18, 2026
About this calculator
The minimum detectable effect (MDE) calculator determines the smallest statistically significant effect a study can identify, given a sample size, significance level (α), and statistical power (1-β). It's critical for experiments like A/B testing or clinical trials to ensure reliable results. The formula used is σ·(z_α + z_β)/√n, where σ represents the sample standard deviation.
The calculation combines critical z-values for α (significance) and β (power) with the standard deviation and square root of the sample size. A smaller MDE indicates higher study sensitivity, while a larger MDE suggests the experiment might miss small effects. Adjusting α, β, or n directly impacts the outcome.
Use this tool before launching an experiment to size the required sample or validate the minimum relevant effect hypothesis. It helps prevent false negatives (β) or false positives (α). Note: Theoretical values might vary in real-world contexts due to unaccounted variability or sampling biases.
Frequently asked questions
How to choose α and β values?
Use α = 0.05 for standard tests and β = 0.2 (80% power) as a reference. Stricter values (e.g., α = 0.01, β = 0.1) increase reliability but require larger samples.
What if the calculated MDE is too high?
A high MDE suggests the study might miss small effects. Reduce it by increasing the sample size or accepting a higher β risk.
Does the calculator work for one-tailed tests?
Yes, but adjust z_α and z_β values according to the directional alternative hypothesis. Two-tailed tests use z_α/2.
Do I need the exact population standard deviation?
During planning, use historical data or pilot studies. For post-hoc analyses, use the sample's observed standard deviation.