Choose Index below for a list of all words and phrases defined in this glossary.

# Type I Error

index | Index

## Type I Error - definitions

Type I Error - An incorrect decision to reject something (such as a statistical hypothesis or a lot of products) when it is acceptable.

[Category=Quality ]

Source: American Society for Quality, 27 October 2010 08:35:19, http://www.asq.org/glossary/

These advertisers support this free service

Type I Error - In hypothesis testing, rejecting the null hypothesis (no difference) when it is in fact true (e.g. convicting an innocent person.)

TYPE 1 errors are those where scientists assumed a relationship where none existed. The Producers risk: Rejecting a good part.

When a point falls out of the boundary limit and the SPC system gives signal that the process is out of control or produced product is bad in Quality_ but actually nothing have gone wrong (i.e., the process is in control).

[Category=Data Quality ]

Source: iSixSigma, 01 March 2011 08:48:04, https:web.archive.org/web/20111109014246/http:www.isixsigma.com/index.php?option=com_glossary

Type I & Type II Errors - Type I error (also known as alpha error) - conclude a difference exists when no difference exists. (for example, you say two machines produce different mean outputs when they do not.).

Type II error (also known as beta error) - conclude no difference exists when it does. (for example, say two machines produce similar mean outputs when in fact they do).

Notes:

a) for fixed sample size experiments, reducing Type I errors result in higher Type II errors. (and vice versa) b) increase in sample size (n), generally reduces both types of errors c) very large sample sizes may result in detecting "statistically significant, but practically insignificant results".

To determine if something is statistically significant, we typically calculate a p-value. To determine statistical significance - a) if p-value is <= alpha, conclude statistical difference, b) if p-value is > alpha, fail to conclude difference. For most experiments: let alpha = 0.01 or 0.05; may tighten alpha if effect of Type I Error is very severe. In terms of statistical significance, (1-p) represents your confidence that a statistically significant difference exists.

[Category=Quality ]

Source: The Quality Portal, 28 April 2011 08:51:50, http://thequalityportal.com/glossary/s.htm

Data Quality Glossary.  A free resource from GRC Data Intelligence. For comments, questions or feedback: dqglossary@grcdi.nl