Statistical Power Interview Questions
Statistical power interview questions for detecting true effects, sample size, variance, effect size, and experiment tradeoffs.
Candidates preparing for experiments and hypothesis-testing prompts.
Power is detection probability
Statistical power is the probability of rejecting the null when a specified alternative is true, not a guarantee that every real effect will be found.
Power depends on effect and noise
Larger effects, lower variance, larger samples, and less strict thresholds can increase power, though each has tradeoffs.
Concrete example
A tiny effect with noisy outcomes may require a very large sample to detect reliably. The issue is signal relative to noise.
Connect to sample size
Power questions often become sample-size questions. Explain how more observations reduce standard error under the usual assumptions.
Common mistakes
Candidates often discuss p-values without mentioning power. A test can fail to find an effect because the design was underpowered.
Practice the pattern
Use the LeetQuidity curriculum and calibration to turn this topic into a focused practice plan.