Bonferroni α ajustado
- Created by
- Renato Passos, Eng. de Software
- Reviewed by
- Renato Passos, Eng. de Software
Last updated: Apr 18, 2026
About this calculator
The Bonferroni adjusted α calculator controls the family-wise error rate (FWER) in multiple comparisons after an ANOVA. It divides the original significance level (α) by the number of comparisons (k), reducing false positives. For example, if α = 0.05 and there are 4 comparisons, the adjusted α becomes 0.0125. This maintains statistical reliability when testing multiple groups in experiments.
To calculate, input the original α and the number of comparisons. The formula is adjusted α = α / k. This conservative approach is ideal for minimizing Type I errors (false positives) but may reduce test power, increasing Type II errors (false negatives) with many comparisons.
Use this method in studies with up to 5-10 comparisons, such as post-ANOVA group comparisons. It's common in biology, psychology, and social sciences. Avoid it for hundreds of tests; alternatives like Holm or False Discovery Rate (FDR) are more efficient. Always validate data with graphical visualizations before applying corrections.
Note: Bonferroni can be overly strict in some cases. If the adjusted α is too low (e.g., 0.001), consider less conservative tests. Also verify comparisons are independent and the initial ANOVA was significant. No method perfectly corrects all scenarios, so combine with qualitative analysis of results.
Frequently asked questions
Why use Bonferroni instead of other methods?
Bonferroni is ideal for few comparisons when avoiding false positives is critical. Methods like Tukey or Holm offer greater statistical power in specific scenarios.
How does the calculator handle non-integer comparison counts?
The calculator requires k to be a positive integer, as it represents real group or pair comparisons to be tested.
Can this method work with non-standard α values like 0.1?
Yes, any value between 0 and 1 is accepted. The result will be α/k, even if the value is unconventional for scientific studies.
Can I use this calculation without doing an ANOVA first?
No. Bonferroni is a post-ANOVA technique. Multiple tests without initial ANOVA lack valid statistical foundation.
What's the maximum number of comparisons allowed?
There's no technical limit, but above 20 comparisons, methods like FDR are more appropriate to avoid excessive power reduction.