Differential Item Functioning For Age Research Articles

Assessment of Differential Item Functioning for Demographic Comparisons in the MOS SF-36 Health Survey

To investigate whether items of the Medical Outcomes Study (MOS) 36-Item Short-Form Health Status Survey (SF-36) exhibited differential item functioning (DIF) with respect to age, education, race, and gender. The data for this study come from two large national datasets, the MOS and the 1990 National Survey of Functional Health Status (NSFHS). We used logistic regression to identify items exhibiting DIF. We found DIF to be most problematic for age comparisons. Items flagged for age DIF were vigorous activities, bend/kneel/stoop, bathing or dressing, limited in kind of work, health in general, get sick easier than others, expect health to get worse, felt calm and peaceful, and all four vitality items. Items flagged for education DIF include vigorous activities, health in general, health is excellent, felt calm and peaceful, and been a happy person. Vigorous activities, walk more than a mile, health in general, and expect health to get worse were identified as DIF when comparing African-Americans with whites. No items were identified for gender DIF. We found several consistent patterns of DIF using two national datasets with different population characteristics. In the current study, the effect of DIF rarely transferred to the scale level. Further research is needed to corroborate these results and determine qualitatively why DIF may occur for these specific items.

The Geriatric Depression Scale should be shortened: results of Rasch analysis

The purpose of this study was to examine the unidimensionality, item fit, redundancy and differential item functioning (DIF) of the 15-item version of the Geriatric Depression Scale (GDS) in a community sample of 300 Hong Kong Chinese patients with pneumoconiosis. Participants were randomly selected from the case register of the Pneumoconiosis Compensation Fund Board of Hong Kong. A trained research assistant administered the GDS to all participants. A psychiatrist, who was blind to the GDS scores, conducted a structured clinical interview to diagnose depressive disorders according to the Diagnostic and Statistical Manual for Mental Disorders, Version IV (DSM-IV) criteria. Of the 300 participants, 37 (12.3%) had a DSM-IV diagnosis of depressive disorders. Eleven out of 15 items (73.3%) had INFIT/OUTFIT statistics between 0.7-1.3. Abbreviated versions were created by removal of misfit and redundant items resulting in similar overall performance as the original 15-item GDS. None of the items had significant DIF for age, level of education and cognitive impairment. Although the GDS was overall unidimensional, there was evidence of item redundancy indicating that a shortened version would be as adequate as the original version.