Probability & Statistics Codexery

P-value

Last updated

Under the assumption that the null hypothesis is correct, a p-value indicates the chance of seeing test results that are at least as extreme as those actually observed.[1] The p-value functions as a decision-making instrument for rejecting the null hypothesis, rather than serving as an indicator of a hypothesis's likelihood of being true.[2]

Quick Facts

Field
Statistics
Associated organization
American Statistical Association (ASA)
Related concept
Null hypothesis

Facts from the source article.

Key facts and history

P-value computations date back to the 1700s, when John Arbuthnot studied London birth records from 1629 to 1710.[3] Arbuthnot found that in every year, more males than females were born, and calculated the probability of this outcome as about 1 in 4,836,000,000,000,000,000,000,000.[4] He concluded this was due to divine providence, not chance, rejecting the null hypothesis of equal male and female births.[4] Karl Pearson first formally introduced the p-value in his chi-squared test, using the chi-squared distribution and notating it as capital P.[5]

Fisher proposed the level p = 0.05, or a 1 in 20 chance, as a limit for statistical significance.[6] The 0.05 value was originally proposed by Fisher in 1925.[7] In 2016, the American Statistical Association made a formal statement that p-values do not measure the probability that the studied hypothesis is true.[8] The p-value's widespread adoption in statistics is credited to Ronald Fisher, who featured it prominently in his 1925 publication "Statistical Methods for Research Workers".[6]

An ASA task force in 2019 determined that when p-values and significance tests are used correctly, they enhance the strictness of data-driven conclusions.[8]

What the research sources say

A common criticism is that p-values can only provide evidence against the null-hypothesis, but can never demonstrate that the null-hypothesis is true.[9] Bayesians argue that scientists misinterpret p-values as estimates of the probability that a hypothesis is true, when they quantify merely the probability that the data could have been produced without an effect.[9] Some statisticians suggest adopting S-values (surprisals) instead of p-values, relating statistical relevance to the number of consecutive heads when flipping an unbiased coin.[10] Within the Fisherian framework, p-values are interpreted as an indicator of how strongly the data contradict the null-hypothesis.[9]

References

  1. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: "Not Even Scientists Can Easily Explain P-values", FiveThirtyEight, 2015-11-24; "The ASA's Statement on p-Values: Context, Process, and Purpose", The American Statistician, 7 March 2016, p. 129–133 (original not read by us) en.wikipedia.org
  2. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: "Scientific method: statistical errors", Nature, February 2014, p. 150–152 (original not read by us) en.wikipedia.org
  3. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: "The Descent of Human Sex Ratio at Birth", Springer Science & Business Media, 2007, p. 1–25; "An argument for Divine Providence, taken from the constant regularity observed in the births of both sexes", Philosophical Transactions of the Royal Society of London, 1710, p. 186–190; "Practical Nonparametric Statistics", Wiley, 1999, p. 157–176; "Applied Nonparametric Statistical Methods", Chapman & Hall, 1989; "The History of Statistics: The Measurement of Uncertainty Before 1900", Harvard University Press, 1986, p. 225–226 (we also read that sourc en.wikipedia.org
  4. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: "Statisticians of the Centuries", Springer, 2001, p. 39–42; "A History of Mathematical Statistics from 1750 to 1930", Wiley, 1998, p. 65; "Practical Nonparametric Statistics", Wiley, 1999, p. 157–176 (we also read that source) en.wikipedia.org
  5. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: "On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling", Philosophical Magazin; "On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling", Philosophical Magazin (original not read by us) en.wikipedia.org
  6. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: Biau, "P Value and the Theory of Hypothesis Testing: An Explanation for New Researchers", Clinical Orthopaedics and Related Research, 2010, p. 885–892; Brereton, "P values and multivariate distributions: Non-orthogonal terms in regression models", Chemometrics and Intelligent Laboratory Systems, 2021; "Confusion Over Measures of Evidence (p′s) Versus Errors (α′s) in Classical Statistical Testing", The American Statistician, 2003, p. 171–178 [p. 171]; Fisher 1925, p. 47; Dallal 2012, p. Note 31: [http://www.jerrydallal.com/LHSP/p05.htm Why P=0.0 en.wikipedia.org
  7. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: Fisher, "Breakthroughs in Statistics: Methodology and Distribution", Springer, 1992, p. 66–70 (we also read that source) en.wikipedia.org
  8. Wikipedia, "P-value" (CC BY-SA 4.0), which cites: Wasserstein, "The ASA Statement on p -Values: Context, Process, and Purpose", The American Statistician, 2016-04-02, p. 129–133; Benjamini, "ASA President's Task Force Statement on Statistical Significance and Replicability", Chance, 2021-10-02, p. 10–11 (original not read by us) en.wikipedia.org
  9. "Replacing p-values with Bayes-Factors: A Miracle Cure for the Replicability Crisis in Psychological Science", Replicability-Index, 30 April 2015. facts only, no text reused. replicationindex.wordpress.com
  10. Rovetta, "S-values and Surprisal intervals to Replace P-values and Confidence Intervals: Accepted - January 2024", REVSTAT-Statistical Journal, 2024-01-12. facts only, no text reused. revstat.ine.pt

Each numbered sentence above was checked against the source it cites before publication. Sentences citing Wikipedia adapt text from the English Wikipedia article P-value (contributors), used under CC BY-SA 4.0; this entry is available under the same licence. Other sources were used for facts only; no text or images are copied from them.

More in Probability & Statistics

Sources

Compiled from Wikipedia and the sources listed below. Text from Wikipedia is available under CC BY-SA 4.0; this entry is adapted from it.

Spotted an error? Know more?

Reader corrections go straight into our review queue. Suggest an edit · How this site is sourced

Comments

Loading…
Open in the interactive codex →