• A O Adebiyi.
• Clinical Epidemiology Unit, Department of Community Medicine, University College Hospital, Ibadan, Nigeria.
• Afr J Med Med Sci. 2010 Sep 1; 39 (3): 241-8.

AbstractThe results of many research findings have come under scrutiny in recent years due to the introduction of systematic errors at one stage or the other of the research. Over the years, literature has been rife about the issue of bias with many authors describing unique types of bias. More often, researchers have often been left in the dark about the basic concept of this important phenomenon. Using a method of cross referencing, exploded tree search, consultation of textbooks of epidemiology and conference proceedings, this article examines the basic concept of bias and the current understanding of bias as present in various literature. A simple classification of biases into conceptualization, selection and information biases is proposed. A distinction is made of confounding as describing an association that is true but potentially misleading and bias which on the other hand creates an association that is not true. The article further goes on to describe the different types of biases applicable to different study designs before concluding on the need for researchers to pay attention to the issue of bias so as to make their studies useful to readers.

10 keywords...

hide…