Response bias in below-chance performance: Computation of the parametric measure β

We argue that conceptual problems arise with the parametric measure β when above- and below-chance data are aggregated or compared. Depending on the interpretation of β as a likelihood-ratio measure or an indicator of strictness or bias toward signal or noise, the original formula for β should be retained or modified. The response bias measure β can be retained only if it is interpreted as a likelihood-ratio measure. If it is interpreted as an indicator of strictness or bias toward signal or noise, the original formula has to be modified. One possible modified formula is suggested here. (PsycINFO Database Record (c) 2016 APA, all rights reserved)

Accession Number: 1990-00204-001. Partial author list: First Author & Affiliation: Waldmann, Michael R.; Johann Wolfgang Goethe-U Frankfurt, Inst für Psychologie, Germany. Other Publishers: Psychological Review Company; The Macmillan Company; The Review Publishing Company. Release Date: 20060710. Correction Date: 20161024. Publication Type: Journal (0100), Peer Reviewed Journal (0110). Format Covered: Print. Document Type: Journal Article. Language: English. Major Descriptor: Parametric Statistical Tests; Response Bias; Statistical Measurement. Classification: Statistics & Mathematics (2240). Population: Human (10). References Available: Y. Page Count: 3. Issue Publication Date: Sep, 1989. Publication History: Accepted Date: Mar 9, 1989; Revised Date: Feb 27, 1989; First Submitted Date: Jun 6, 1988. Copyright Statement: American Psychological Association. 1989.