P-Hacking
Definition
Trying multiple analytic choices on the same dataset until one combination crosses a statistical significance threshold, then reporting only that combination as the planned analysis.
Related terms
- Fabrication
- Inventing data or results that were never actually collected, so the reported study never took place.
- Falsification
- Manipulating research materials, equipment or processes, or changing or omitting data, so that the record no longer accurately reflects what happened.
- HARKing
- Hypothesising after the results are known: presenting a hypothesis formed after seeing the data as though it had been predicted in advance.
- Plagiarism
- Presenting another person's words, ideas, data or findings as one's own without adequate credit, including copying one's own previously published work without...
- Retraction
- A formal withdrawal of a published article by a journal or its authors, issued when the findings can no longer be trusted...
- Pre-Registration
- Filing a study's hypotheses, sample size and planned analysis in a public registry before collecting or examining outcome data.
- Questionable Research Practices
- Practices such as p-hacking, HARKing and salami slicing that distort what a study appears to show without meeting the FFP definition of...
Explained in these topics
- Researcher Fraud and Plagiarism (UGC-NET Unit VI)
- Research Misconduct: Fabrication, Falsification and PlagiarismTrying multiple analyses, subgroups or cutoffs on the same dataset until a statistically significant result appears, then reporting only that result.