Skip to main content
DigiCalcs

learn.howToCalculate

learn.whatIsHeading

Determines if A/B test results are statistically significant or due to chance. Critical for making confident product and marketing decisions.

단계별 가이드

  1. 1Define null hypothesis and significance level (typically 0.05)
  2. 2Calculate sample size needed for power
  3. 3Run A/B test long enough
  4. 4Calculate p-value using chi-square or t-test
  5. 5Accept or reject hypothesis

풀어진 예시

입력
A 10%, B 12%
결과
95% confidence

피해야 할 일반적인 실수

  • Stopping test too early when results favor variant
  • Multiple testing inflating false positive rate

자주 묻는 질문

What sample size do I need?

Depends on baseline rate and expected lift; typically 500-2,000 per variant for 80-90% statistical power.

What's statistical significance?

95% confidence the result wasn't due to chance; p-value < 0.05 typically required.

설정