What is the Bonferroni correction, and what is the cost of using it?
PICTURE THIS: DATA SPLIT
Fit on train, tune on val, report on test once.
Simple meaning
Bonferroni splits alpha across m tests by using alpha/m for each, controlling the family-wise error rate under weak assumptions.
WHY — Hypothesis Testing instead of guessing?
Why interviewers care about Hypothesis Testing:
people who only read docs from people who shipped.
and tied to Data Science work.
Name the idea, why it exists, then one short example.
End with when you use it and one common pitfall.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1Bonferroni splits alpha across
m tests by using alpha/m for each, controlling the family-wise error rate under weak assumptions.
- 2It is conservative, especially
with correlated metrics, so you lose power and miss real effects.
- 3Prefer it when a
single false claim among many would be costly, and consider Holm or FDR when many tests are exploratory.
- 4Give an example
One tiny concrete case you can say aloud.
- 5Common mistake
What juniors usually get wrong.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
Bonferroni splits alpha across m tests by using alpha/m for each, controlling the family-wise error rate under weak assumptions. It is conservative, especially with correlated metrics, so you lose power and miss real effects.