Selection Bias
Definition
Selection bias occurs when the process of selecting study participants produces a sample that doesn’t accurately represent the target population. This bias can distort results and limit the validity of conclusions.
Correct Scientific Usage
Researchers use randomization, careful recruitment strategies, and representative sampling to minimize selection bias. When bias is unavoidable, they acknowledge it as a limitation.
Common sources include convenience sampling (recruiting whoever is easiest to access), volunteer bias (self-selected participants may differ from non-volunteers), and survival bias (studying only those who survive or remain in treatment).
Common Misunderstandings
People often don’t realize how selection bias shapes conclusions. For example, surveys sent to existing customers naturally exclude those who stopped using a product—often the very people with the most informative experiences.
There’s also confusion between selection bias and other biases. Selection bias occurs during recruitment; it’s different from response bias (how people answer questions) or publication bias (which studies get published).
Why It Matters
Selection bias can make treatments appear more effective or safer than they are in typical patients. Recognizing it helps explain why results observed in trials may not match outcomes seen in routine clinical practice.
References
- Clinimetrics corner: the many faces of selection bias, Journal of Manual & Manipulative Therapy
- A framework for understanding selection bias in real-world healthcare data, Statistics in Society
Related Terms
Related Articles
- Why Most Health Studies Don't Apply Equally to Everyone
- What Statistical Significance Actually Tells Us
- What ‘Backed by Science’ Really Means
- Why One Study is Almost Never Enough