Premium
Mixed group validation: a method to address the limitations of criterion group validation in research on malingering detection *
Author(s) -
Frederick Richard I.
Publication year - 2000
Publication title -
behavioral sciences and the law
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.649
H-Index - 74
eISSN - 1099-0798
pISSN - 0735-3936
DOI - 10.1002/bsl.432
Subject(s) - malingering , psychology , lie detection , test (biology) , group (periodic table) , statistics , social psychology , mathematics , clinical psychology , deception , biology , physics , paleontology , quantum mechanics
Mixed group validation (MGV) is offered as an alternative to criterion group validation (CGV) to estimate the true positive and false positive rates of tests and other diagnostic signs. CGV requires perfect confidence about each research participant's status with respect to the presence or absence of pathology. MGV determines diagnostic efficiencies based on group data; knowing an individual's status with respect to pathology is not required. MGV can use relatively weak indicators to validate better diagnostic signs, whereas CGV requires perfect diagnostic signs to avoid error in computing true positive and false positive rates. The process of MGV is explained, and a computer simulation demonstrates the soundness of the procedure. MGV of the Rey 15‐Item Memory Test (Rey, 1958) for 723 pre‐trial criminal defendants resulted in higher estimates of true positive rates and lower estimates of false positive rates as compared with prior research conducted with CGV. The author demonstrates how MGV addresses all the criticisms Rogers (1997b) outlined for differential prevalence designs in malingering detection research. Copyright © 2000 John Wiley & Sons, Ltd.