Contact

Tidbits from Stats Calculator Online

• Best practice

  Check assumptions before you trust p

  Independence, normality (or large n), and equal variances matter for classic t/ANOVA. Use Welch or rank tests when they fail.

  Two-sample t-test →

• Plan ahead

  Power before you collect data

  Fix the smallest meaningful effect and α, then compute sample size. Inflate n for dropout when recruiting humans.

  Sample size calculator →

• Everyday math

  Percent change ≠ percent of

  Percent of is part÷whole. Percent change uses the starting value in the denominator—and +50% then −50% does not return to start.

  Percentage calculators →