| Bonferroni |
One-step correction |
Use when the number of comparisons is small or when strict control of the family-wise error rate is needed. |
| Sidak |
One-step correction |
Similar to Bonferroni, but slightly less conservative. Use when controlling the family-wise error rate. |
| Holm-Sidak |
Step-down method using Sidak adjustments |
Use when controlling the family-wise error rate with a step-down approach. |
| Holm |
Step-down method using Bonferroni adjustments |
Use when controlling the family-wise error rate with a step-down approach. |
| Simes-Hochberg |
Step-up method (independent) |
Use when controlling the false discovery rate (FDR) with independent tests. |
| Hommel |
Closed method based on Simes tests (non-negative) |
Use when controlling the FDR with non-negative dependencies among tests. |
| Benjamini/Hochberg |
Benjamini/Hochberg (non-negative) |
Use when controlling the FDR and the tests are independent or positively correlated. |
| Benjamini/Yekutieli |
Benjamini/Yekutieli (negative) |
Use when controlling the FDR and the tests are dependent, possibly negatively correlated. |
| Two-stage Benjamini/Hochberg |
Two-stage FDR correction (non-negative) |
Use when controlling the FDR with potentially dependent tests, particularly for non-negative dependencies. |
| Two-stage Benjamini/Yekutieli |
Two-stage FDR correction (non-negative) |
Similar to Two-stage Benjamini/Hochberg but offers improved control over the FDR under certain conditions. |