Related Experiment Videos
Confounded by confounding: separating association from causation
1Department of Management, University of St Andrews, Fife.
Hospital Medicine (London, England : 1998)
|July 9, 1999
Abstract:
Health-care professionals need to be able to distinguish causal relationships from simple associations in two main areas: when unravelling the aetiology of diseases, and when assessing the effects of therapies. In each of these the presence of confounding can seriously mislead. This short report explains the nature of confounding and outlines criteria that can be applied to help distinguish causality from mere statistical associations.