What are always-valid p-values or sequential testing methods used for in online experiments?
PICTURE THIS: FLEX VS GRID
Simple meaning
They let you look at results continuously while still controlling Type I error, using martingale or mixture sequential probability ratio ideas.
WHY — P-value instead of guessing?
Why interviewers care about P-value:
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:
- 1They let you look
at results continuously while still controlling Type I error, using martingale or mixture sequential probability ratio ideas.
- 2That supports peeking and
early stopping without the naive t-test penalty.
- 3They typically need more
data for the same MDE than a fixed-horizon test, which is the price of flexibility.
- 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
They let you look at results continuously while still controlling Type I error, using martingale or mixture sequential probability ratio ideas. That supports peeking and early stopping without the naive t-test penalty.