A Dynamic Thurstonian Item Response Theory of Motive Expression in the Picture Story Exercise: Solving the Internal Consistency Paradox of the PSE

The measurement of implicit or unconscious motives using the picture story exercise (PSE) has long been a target of debate in the psychological literature. Most debates have centered on the apparent paradox that PSE measures of implicit motives typically show low internal consistency reliability on...

Full description

Saved in:
Bibliographic Details
Published in:Psychological review 2014-07, Vol.121 (3), p.481-500
Main Author: Lang, Jonas W. B.
Format: Article
Language:English
Subjects:
Biological and medical sciences
Choice Behavior
Choice Behavior - physiology
Cognition. Intelligence
Computer Simulation
Correlation analysis
Decision making models
Decision making. Choice
Estimation bias
Forecasts
Fundamental and applied biological sciences. Psychology
Human
Humans
Internal Consistency
Item Response Theory
Mathematical models
Measurement
Models, Psychological
Motivation
Motivation - physiology
Pictorial Stimuli
Psychological Theory
Psychology
Psychology. Psychoanalysis. Psychiatry
Psychology. Psychophysiology
Psychometrics
Quantitative psychology
Reliability
Simulation
Storytelling
Test Reliability
Written Communication
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:The measurement of implicit or unconscious motives using the picture story exercise (PSE) has long been a target of debate in the psychological literature. Most debates have centered on the apparent paradox that PSE measures of implicit motives typically show low internal consistency reliability on common indices like Cronbach's alpha but nevertheless predict behavioral outcomes. I describe a dynamic Thurstonian item response theory (IRT) model that builds on dynamic system theories of motivation, theorizing on the PSE response process, and recent advancements in Thurstonian IRT modeling of choice data. To assess the models' capability to explain the internal consistency paradox, I first fitted the model to archival data (Gurin, Veroff, & Feld, 1957) and then simulated data based on bias-corrected model estimates from the real data. Simulation results revealed that the average squared correlation reliability for the motives in the Thurstonian IRT model was .74 and that Cronbach's alpha values were similar to the real data (
ISSN:0033-295X
1939-1471
DOI:10.1037/a0037011