HARKing (Hypothesizing After Results Are Known) is generally regarded as a deceitful practice in science. The statistical tests associated with hypothesis testing are based on assumptions of a prioroi hypothesis, and a limited number to be tested. Once these are violated, so are the validity of the tests; and probability of false results increase dramatically.

This editorial lays out the case against HARKing, when done in secret (SHARKing); discusses external forces that drive the practice (such as helpful suggestions from editors and reviewers) and makes the case for exploring quantitative data transparently (THARKing); so as to both get as much value from the data as possible and lay the groundwork for future studies.


Share This