Systematic selection bias: A cause of dramatic errors in the inference of treatment effectiveness

Abstract

Our understanding of the world around us is derived from our observations. The accuracy of our inferences depends on the representativeness of those observations. Selection bias limits the accuracy of our inferences. Systematic selection bias occurs when the particular outcome observed caused the observation; this type of bias can lead to dramatic errors in inference. We describe examples of selection bias, provide a mathematical formulation of the systematic selection bias phenomenon, and discuss how biased observations may affect people's impressions of important issues in dermatology.

• Epidemiology
• health care delivery
• health services research
• office surgery
• regulations