Abstract
In this review I trace the origins, applications, limitations and future prospects for research on measurement item bias, or differential item functioning (DIF) in the context of health research. DIF arises in the context of using multiple item or symptom health instruments to rate the level of a particular condition, and describes the situation where not all persons at the same level at the same level of the underlying condition have the same probability of endorsing one or more symptoms. The presence of DIF can lead to biased assessment of group differences and confound risk factor and outcomes research. The epidemiologic literature includes a great many applied, review, and methodological articles focusing on DIF. The preponderance of the literature appears in the areas of health-related quality of life, physical functioning, cognition, and mental health outcomes. Epidemiologists and other researchers in the health sciences often rely upon multiple item rating scales or questionnaires to assess for the presence of or level of health conditions or states that are otherwise not directly observable. When population subgroups respond differently to a subset of the items, this is referred to as differential item functioning (DIF), and might be a source of bias.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have