The validity of statistical findings can always be called into question due to 1) the biases (selection and observation) that exist in research methodologies and 2) the assumptions associated with conducting inferential statistics (normality, homoscedasticity, linearity). There are three primary methods for validating statistical findings: Split-group, jack-knife, and bootstrap. Click on a button below to learn more about validating statistical findings.
Take thousands of random samples from a group of participants to yield 95% confidence intervals for statistical differences or associations.
Randomly assign participants into one of two groups, the derivation group and the confirmatory group, and run statistics on each set.
Run statistical analyses on every independent sample possible with each participant removed one time.
