Going beyond the detection of differential item functioning in tests for personnel selection
Differential item functioning (DIF) occurs when persons with equal ability who belong to different groups have a different probability to correctly solve an item of a certain test. As the occurrence of DIF is considered a serious problem when using tests for personnel selection, many procedures have been developed for detecting DIF. However, models for explaining DIF in a systematic manner have received only little attention. This paper discusses a further extension of item response theory based models that provides a strong substantive basis for explaining DIF. The general idea is to model item difficulties in each group as a function of item features such as the cognitive processes that are needed for correctly solving items. A different weight for a particular feature across groups also called differential feature functioning then provides a strong substantive basis for explaining why DIF may occur in items that share this particular feature. Differential feature functioning can be used for estimating the average difference between feature weights across groups. Another useful model extension is to consider feature weights that are random over persons. Models with random feature weights allow to assess the extent to which feature weights tend to differ between groups as well as within groups. As an illustration of the approach, the proposed models will be used for modeling DIF in a test for transitive reasoning using data from personnel selection.
|Date of creation:||Dec 2009|
|Date of revision:|
|Contact details of provider:|| Web page: http://research.hubrussel.be|
More information through EDIRC
References listed on IDEAS
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- David J. Spiegelhalter & Nicola G. Best & Bradley P. Carlin & Angelika van der Linde, 2002. "Bayesian measures of model complexity and fit," Journal of the Royal Statistical Society Series B, Royal Statistical Society, vol. 64(4), pages 583-639.
- Nambury Raju, 1988. "The area between two item characteristic curves," Psychometrika, Springer, vol. 53(4), pages 495-502, December.
- Henk Kelderman, 1989. "Item bias detection using loglinear irt," Psychometrika, Springer, vol. 54(4), pages 681-697, September.
When requesting a correction, please mention this item's handle: RePEc:hub:wpecon:200940. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Sabine Janssens)
If references are entirely missing, you can add them using this form.