Selection Bias and Its Causes
What you'll learn: How choosing samples in a non-random way can lead to misleading conclusions that don't reflect reality.
What Is Selection Bias?
Selection bias occurs when the way you select your sample systematically excludes or overrepresents certain groups from your population. Unlike random sampling methods (simple random, stratified, cluster), selection bias happens when something in your sampling process favors particular outcomes, making your sample unrepresentative.
Think of it like trying to understand what "typical Americans" think by only asking people at a yacht club. You'd get wealthy perspectives but miss everyone else—your conclusions would be systematically skewed.
Common Causes of Selection Bias
Convenience sampling gone wrong: Surveying only people easy to reach (like students in your class) when you want to understand all adults.
Self-selection: When people choose whether to participate. Online polls suffer from this—only people with strong opinions bother responding.
Survivorship bias: Only studying "survivors" or successes. If you analyze only companies that succeeded, you miss lessons from the failures.
Non-response bias: When certain groups don't respond to your survey. If young people ignore phone surveys, your results skew toward older respondents.
Undercoverage: Your sampling frame (the list you sample from) doesn't include part of the population. Phone surveys miss people without phones.
Why It Matters
Selection bias makes your sample systematically different from your population. This isn't random error—it pushes your estimates in one direction consistently. No matter how large your sample gets, it stays biased. You can't fix it with more data if the selection process itself is flawed.
Key Takeaway: Selection bias arises when your sampling method systematically excludes or overrepresents parts of the population, leading to conclusions that don't generalize—and unlike random error, it doesn't decrease with larger samples.