Eliciting response bias within forced choice tests to detect random responders

Robin Orthey, Aldert Vrij, Ewout Meijer, Sharon Leal, Hartmut Blank

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

The Forced Choice Test (FCT) can be used to detect malingered loss of memory or sensory deficits. In this test, examinees are presented with two stimuli, one correct and one incorrect, in regards to a specific event or a perceptual discrimination task. The task is to select the correct answer alternative, or guess if it is unknown. Genuine impairment is associated with test scores that fall within chance performance. In contrast, malingered impairment is associated with purposeful avoidance of correct information, resulting in below chance performance. However, a substantial proportion of malingerers intentionally randomize their responses, and are missed by the test. Here we examine whether a 'runs test' and a 'within test response 'bias' have diagnostic value to detect this intentional randomization. We instructed 73 examinees to malinger red/green blindness and subjected them to a FCT. For half of the examinees we manipulated the ambiguity between answer alternatives over the test trials in order to elicit a response bias. Compared to a sample of 10,000 cases of computer generated genuine performance, the runs test and response bias both detected malingered performance better than chance.

Original languageEnglish
Article number8724
Number of pages7
JournalScientific Reports
Volume9
Issue number1
DOIs
Publication statusPublished - 19 Jun 2019

Cite this

@article{e560351efd424efe8b007595196e6b10,
title = "Eliciting response bias within forced choice tests to detect random responders",
abstract = "The Forced Choice Test (FCT) can be used to detect malingered loss of memory or sensory deficits. In this test, examinees are presented with two stimuli, one correct and one incorrect, in regards to a specific event or a perceptual discrimination task. The task is to select the correct answer alternative, or guess if it is unknown. Genuine impairment is associated with test scores that fall within chance performance. In contrast, malingered impairment is associated with purposeful avoidance of correct information, resulting in below chance performance. However, a substantial proportion of malingerers intentionally randomize their responses, and are missed by the test. Here we examine whether a 'runs test' and a 'within test response 'bias' have diagnostic value to detect this intentional randomization. We instructed 73 examinees to malinger red/green blindness and subjected them to a FCT. For half of the examinees we manipulated the ambiguity between answer alternatives over the test trials in order to elicit a response bias. Compared to a sample of 10,000 cases of computer generated genuine performance, the runs test and response bias both detected malingered performance better than chance.",
author = "Robin Orthey and Aldert Vrij and Ewout Meijer and Sharon Leal and Hartmut Blank",
year = "2019",
month = "6",
day = "19",
doi = "10.1038/s41598-019-45292-y",
language = "English",
volume = "9",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",
number = "1",

}

Eliciting response bias within forced choice tests to detect random responders. / Orthey, Robin; Vrij, Aldert; Meijer, Ewout; Leal, Sharon; Blank, Hartmut.

In: Scientific Reports, Vol. 9, No. 1, 8724, 19.06.2019.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Eliciting response bias within forced choice tests to detect random responders

AU - Orthey, Robin

AU - Vrij, Aldert

AU - Meijer, Ewout

AU - Leal, Sharon

AU - Blank, Hartmut

PY - 2019/6/19

Y1 - 2019/6/19

N2 - The Forced Choice Test (FCT) can be used to detect malingered loss of memory or sensory deficits. In this test, examinees are presented with two stimuli, one correct and one incorrect, in regards to a specific event or a perceptual discrimination task. The task is to select the correct answer alternative, or guess if it is unknown. Genuine impairment is associated with test scores that fall within chance performance. In contrast, malingered impairment is associated with purposeful avoidance of correct information, resulting in below chance performance. However, a substantial proportion of malingerers intentionally randomize their responses, and are missed by the test. Here we examine whether a 'runs test' and a 'within test response 'bias' have diagnostic value to detect this intentional randomization. We instructed 73 examinees to malinger red/green blindness and subjected them to a FCT. For half of the examinees we manipulated the ambiguity between answer alternatives over the test trials in order to elicit a response bias. Compared to a sample of 10,000 cases of computer generated genuine performance, the runs test and response bias both detected malingered performance better than chance.

AB - The Forced Choice Test (FCT) can be used to detect malingered loss of memory or sensory deficits. In this test, examinees are presented with two stimuli, one correct and one incorrect, in regards to a specific event or a perceptual discrimination task. The task is to select the correct answer alternative, or guess if it is unknown. Genuine impairment is associated with test scores that fall within chance performance. In contrast, malingered impairment is associated with purposeful avoidance of correct information, resulting in below chance performance. However, a substantial proportion of malingerers intentionally randomize their responses, and are missed by the test. Here we examine whether a 'runs test' and a 'within test response 'bias' have diagnostic value to detect this intentional randomization. We instructed 73 examinees to malinger red/green blindness and subjected them to a FCT. For half of the examinees we manipulated the ambiguity between answer alternatives over the test trials in order to elicit a response bias. Compared to a sample of 10,000 cases of computer generated genuine performance, the runs test and response bias both detected malingered performance better than chance.

U2 - 10.1038/s41598-019-45292-y

DO - 10.1038/s41598-019-45292-y

M3 - Article

C2 - 31217488

VL - 9

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

IS - 1

M1 - 8724

ER -