Examples
Worked examples
- Is an instance
Trying several different outcome definitions, covariate sets, or subgroup splits, then reporting only the combination that reached statistical significance, without disclosing the other analyses that were tried and discarded
Counter-examples
Looks similar, but isn't
- Not an instance
Running one pre-registered analysis exactly as specified in advance, and reporting a null result honestly, is not p-hacking even though the researcher may have hoped for a significant finding — the practice is defined by undisclosed selective reporting, not by the outcome
Editorial commentary
P-hacking is the practice of exploiting flexibility in data collection, exclusion, or analysis choices — sometimes deliberately, more often unconsciously — to steer a result toward statistical significance, without disclosing that other reasonable analytic choices were tried and discarded. It is one of the best-documented questionable research practices and a widely cited contributor to the field’s reproducibility crisis.
How it differs from HARKing
P-hacking and HARKing (Hypothesising After the Results are Known) are related but distinct failure modes: p-hacking is about the analysis — trying multiple statistical approaches until one reaches significance — while HARKing is about the write-up, presenting a post-hoc finding as though it had been the study’s original hypothesis all along. The two often occur together but are separable practices with separate defences. See P-hacking vs HARKing for a direct comparison.
Why it is hard to detect from a single paper
Because the undisclosed alternative analyses never appear in the published report, p-hacking is usually invisible in any one paper. It is instead detected at the literature level: an unusual excess of reported p-values clustered just under the 0.05 significance threshold across many published studies is a documented statistical signature (the “p-curve” method), and discrepancies between a study’s pre-registered analysis plan and what was actually reported are a direct, paper-level red flag.
How journals and funders respond
The principal structural defence is pre-registration of hypotheses and analysis plans before data collection or analysis begins, which removes the researcher’s ability to select an analysis after seeing the results. Registered Reports go further, having a journal peer-review and provisionally accept the study design and analysis plan before results exist at all, removing any incentive to p-hack toward a publishable outcome. Many journals now also require a statistical analysis plan or disclosure of all analyses conducted, not only those reported.
References
- Simmons, Nelson, Simonsohn, ‘False-positive psychology’ (Psychological Science, 2011); Simonsohn, Nelson, Simmons, ‘P-curve’ (Journal of Experimental Psychology: General, 2014).
Also known as
p-fishing · significance chasing
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="P-hacking"
vocab-term-identifier="https://casrai.org/dictionary/term/p-hacking" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/p-hacking",
"name": "P-hacking",
"identifier": "https://casrai.org/dictionary/term/p-hacking",
"description": "P-hacking is the practice of exploiting flexibility in data collection, exclusion, or analysis choices — often unconsciously — to steer a result toward statistical significance, undisclosed in the final report as if the reported analysis were the only one that had been tried.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/reproducibility#set",
"url": "https://casrai.org/dictionary/term/p-hacking",
"sameAs": [
"p-fishing",
"significance chasing"
],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"author": {
"@id": "https://casrai.org/#editorial-team"
},
"datePublished": "2026-05-21T02:22:50",
"dateModified": "2026-08-22T15:39:33",
"inLanguage": "en-GB",
"isAccessibleForFree": true
}
