Detection Theory Analysis of Group Data: Estimating Sensitivity From Average Hit and False-Alarm Rates

This article assesses the statistical consequences of computing detection-theoretic sensitivity measures ( d ′) from response rates averaged across subjects. Collapsed d ′, based on such averaged proportions, underestimates true average d ′ to only a small degree in most cases, and its variability i...

Full description

Saved in:
Bibliographic Details
Published in:Psychological bulletin 1985-07, Vol.98 (1), p.185-199
Main Authors: Macmillan, Neil A, Kaplan, Howard L
Format: Article
Language:English
Subjects:
Biological and medical sciences
Fundamental and applied biological sciences. Psychology
Human
Humans
Perception - physiology
Psychology. Psychoanalysis. Psychiatry
Psychology. Psychophysiology
Psychometrics. Statistics. Methodology
Psychophysics - methods
Statistical Estimation
Statistics as Topic - methods
Statistics. Mathematics
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This article assesses the statistical consequences of computing detection-theoretic sensitivity measures ( d ′) from response rates averaged across subjects. Collapsed d ′, based on such averaged proportions, underestimates true average d ′ to only a small degree in most cases, and its variability is less than that of averaged d ′. We consider two other approaches to computing group sensitivity: transforming the individual response rates before averaging and using nonparametric measures. Careful application of detection theory in situations in which individual sensitivities cannot be computed is often better than these alternatives.
ISSN:0033-2909
1939-1455
DOI:10.1037/0033-2909.98.1.185