Skip to lesson
Exit
Experimentation & Causal Inference1 / 2

1 min lesson

Peeking and sequential testing

Use "Peeking and sequential testing" to tell the cases apart, then choose a response for each one.

Step 1 of 2

Peeking and sequential testingthe bug that fakes wins

A fixed-horizon test promises one significance level if you look once, at the end. Engineers in a hurry look every morning and ship the moment p drops below 0.05. That naive repeated peeking can inflate the false-positive rate past 30 percent. You either commit to a fixed sample and look once or you use a method built for continuous monitoring.

Always-valid p-values

Sequential / mSPRT-style inference

Look as often as you want, type-I error still controlled

Pay for it with a wider effective threshold

Group-sequential designs

Pre-planned interim looks with alpha-spending

Can stop early for a clear win or for futility

Needs the look schedule fixed up front