Abstract
Background: Undetected aberrant responses in questionnaire data cause biases and improper diagnostic decisions. Some questionnaires use algorithms to detect aberrant responses based on predefined criteria, but most questionnaires lack this possibility.
Objective: A method for the identification of aberrant responses based on Rasch analysis, is introduced.
Methods: As the first step a homogeneous scale is built according to the Rasch model, and in the second step persons with aberrant responses are identified based on this scale. The procedure is demonstrated with a psychological assessment instrument (FEP) on the basis of a mixed sample (N = 537).
Conclusion: This procedure generates information useful for clinical-psychological applications in practice.
Objective: A method for the identification of aberrant responses based on Rasch analysis, is introduced.
Methods: As the first step a homogeneous scale is built according to the Rasch model, and in the second step persons with aberrant responses are identified based on this scale. The procedure is demonstrated with a psychological assessment instrument (FEP) on the basis of a mixed sample (N = 537).
Conclusion: This procedure generates information useful for clinical-psychological applications in practice.
| Translated title of the contribution | Identification of aberrant responses in the ”Questionnaire for the evaluation of treatment progress based on Rasch modeling |
|---|---|
| Original language | German |
| Pages (from-to) | 135-144 |
| Number of pages | 10 |
| Journal | Zeitschrift für Klinische Psychologie und Psychotherapie |
| Volume | 38 |
| Issue number | 2 |
| DOIs | |
| Publication status | Published - 21 Apr 2009 |
Keywords
- probabilistic test theory
- Rasch analysis
- aberrant response patterns