Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Type I and type II errors
open-in-new
<p> In <a href="/facts/Statistical_hypothesis_testing/yv9RPF6U">statistical hypothesis testing</a>, a type I error, or a false positive, is the erroneous rejection of a true <a href="/facts/Null_hypothesis/C8NwSE9J">null hypothesis</a>. A type II error, or a false negative, is the erroneous failure in bringing about appropriate rejection of a false null hypothesis. </p><p>Type I errors can be thought of as errors of commission, in which the <a href="/facts/Status_quo/kMsM2X2s">status quo</a> is erroneously rejected in favour of new, misleading information. Type II errors can be thought of as errors of omission, in which a misleading status quo is allowed to remain due to failures in identifying it as such. For example, if the assumption that people are <a href="/facts/Presumption_of_innocence/hGuaDx4L"><i>innocent until proven guilty</i></a> were taken as a null hypothesis, then proving an innocent person as guilty would constitute a Type I error, while failing to prove a guilty person as guilty would constitute a Type II error. If the null hypothesis were inverted, such that people were by default presumed to be <i>guilty until proven innocent</i>, then proving a guilty person's innocence would constitute a Type I error, while failing to prove an innocent person's innocence would constitute a Type II error. The manner in which a null hypothesis frames contextually default expectations influences the specific ways in which type I errors and type II errors manifest, and this varies by context and application. </p><p>Knowledge of type I errors and type II errors is applied widely in fields of in <a href="/facts/Medical_science/2fnp90Je">medical science</a>, <a href="/facts/Biometrics/MgTbyA0r">biometrics</a> and <a href="/facts/Computer_science/lN3pEyPz">computer science</a>. Minimising these errors is an object of study within <a href="/facts/Statistical_theory/7s5ye1mp">statistical theory</a>, though complete elimination of either is impossible when relevant outcomes are not determined by known, observable, causal processes. </p>