Module 12 Checkpoint: Experiments, A/B Tests, and Statistical Reporting
Scenario
An A/B test was checked daily and stopped on day 9, when variant B first reached p < 0.05 on one of five metrics.
Questions
- Name two problems with how this test was run.
- What should have been decided before the test started?
- Why does checking daily and stopping at significance inflate the error rate?
- Five metrics were tracked. How does that affect one significant result?
- Write the honest summary of what this test established.
Pass Standard
You pass when you can name the stopping-rule and multiple-metric problems, and write a summary that does not claim more than the test supports.
