Benjamini-Hochberg FDR Correction
The Benjamini-Hochberg procedure is a multiple testing correction that controls the false discovery rate, the expected proportion of false positives among rejected hypotheses, by comparing ordered p-values with rank-scaled thresholds.
벤야미니-호흐베르크 절차는 순서화된 p값을 순위에 비례한 임계값과 비교하여, 기각된 가설 중 거짓 양성의 기대 비율인 거짓발견률을 통제하는 다중검정 보정 방법입니다.
How do I correct p-values when I run many group comparisons or tests at once?
여러 집단 비교나 검정을 한꺼번에 할 때 p값은 어떻게 보정하나요?
The Benjamini-Hochberg procedure adjusts a set of p-values so that the expected proportion of false discoveries among the significant results stays at a chosen level. It suits studies with many comparisons, such as numerous group contrasts or outcomes, where Bonferroni-type familywise control would be too conservative.
When to use Benjamini-Hochberg FDR Correction
- Many pairwise group comparisons on the same outcome
- Testing the same group difference across many outcomes or variables
- Exploratory screening where some false positives are acceptable
When to use something else
- Any single false positive is costly: use the Bonferroni or Holm correction
- Tests with arbitrary or negative dependence: use the Benjamini-Yekutieli procedure
Assumptions and requirements
- Valid p-values from each individual test
- Tests independent or positively dependent
- Target false discovery rate chosen in advance
What the analysis reports
- Raw and BH-adjusted p-values
- List of hypotheses rejected at the chosen false discovery rate
- Number of discoveries out of the total tests
How to report it
After Benjamini-Hochberg correction, [k] of [m] comparisons remained significant at a false discovery rate of [value].
Related methods
- Bonferroni Correction
- Holm-Bonferroni Method
- Benjamini-Yekutieli Procedure
Run Benjamini-Hochberg FDR Correction on Quorant
- Data you need
- a table (CSV or Excel) with an outcome variable and a grouping variable
- Price
- 150 credits for one analysis up to 10,000 rows (complexity level 1); larger or wider data costs more. Optional methodologist review: 900 credits.
- Checking
- Python results are cross-checked against R reference outputs in every deployment
- Before you pay
- A free pre-analysis diagnosis checks your data and gives an exact quote; nothing is deducted until you approve it.