Sampling Bias: The Intuition
Imagine you want to know the average height of all students in your school. You decide to collect data only from the basketball team. You measure them, calculate the average, and announce: "The average student in this school is 6 feet tall."
Your classmates would laugh. Why? Because you didn't measure a fair cross-section of the school — you measured a group that is systematically taller than the rest. Your sample was biased toward tall people.
That's sampling bias in a nutshell: the sample you collected does not represent the population you want to talk about.
The Precise Statement
Sampling bias occurs when the method used to select a sample causes some members of the population to be systematically more likely to be included than others, so that the sample is not representative of the population.
The key word is systematically. Random chance can give you a weird sample once in a while — that's just bad luck. But sampling bias means your selection process itself is flawed, so every sample you draw will be skewed in the same direction.
Why It Matters
If your sample is biased, every conclusion you draw from it is suspect. The average you compute, the proportion you estimate, the trend you observe — all of them will be shifted away from the true population value. This is not a small error; it's a systematic error that no amount of averaging or large sample size can fix.
A large biased sample is not better than a small biased sample. Making the sample bigger only makes the wrong answer more precise — it doesn't make it right.
Common Types of Sampling Bias
| Type | What happens | Example |
|---|
| Selection bias | Your sampling method excludes certain groups | Phone surveys miss people without phones |
| Survivorship bias | You only observe things that "survived" a process | Studying successful startups ignores failed ones |
| Voluntary response bias | People who choose to respond are different from those who don't | Online polls attract people with strong opinions |
| Convenience bias | You sample whatever is easiest to reach | Surveying only people in your classroom |
A Concrete Example
Suppose a school wants to know if students support a new uniform policy. They put a survey on the school website and announce it during morning assembly. 500 students respond, and 80% say they oppose the policy.
Is this reliable? No. The sample is biased because:
- Only students who visit the website saw the survey
- Students with strong opinions (especially negative ones) are more likely to respond
- Students who skip assembly never heard about it
The true opinion of all 2000 students could be completely different.
The Bottom Line
Always ask: "Who is missing from my sample, and why?" If you can answer that honestly, you can spot sampling bias before it ruins your conclusions.
Sampling bias is not about bad intentions — it's about bad design. The best researchers still fall into it. The difference is they check for it, and they admit when their sample has limits.