Population vs. Sample
We rarely study everything. Instead, we study a part and use it to learn about the whole. This distinction is the foundation of statistical inference.
The Population
The entire group of individuals or observations you want to study. Every single one, no exceptions. It can be finite or infinite.
The Sample
A subset of the population — the group you actually observe and collect data from. A good sample is a miniature portrait of the whole.
Why We Sample
Three reasons we cannot study entire populations:
- Cost — surveying millions is prohibitively expensive
- Time — a full census takes years to complete
- Impossibility — some populations are infinite or measurement is destructive
Random Sampling
Every member of the population has an equal chance of being selected. Like drawing names from a hat blindly. This minimizes bias and is the foundation of inference.
Sampling Methods
Beyond simple random sampling:
- Stratified — sample from each subgroup separately
- Cluster — randomly select groups, study all within
- Systematic — every k-th individual from a list
Sampling Bias
When your sample doesn't represent the population. Convenience sampling, voluntary response, and survivorship bias can all distort results — even with large samples.
Sample Size
Margin of error shrinks with the square root of sample size. Quadrupling your sample only halves the error. Quality matters more than quantity.
Parameter vs Statistic
Parameters describe populations (μ, σ) — fixed but unknown. Statistics describe samples (x̄, s) — calculated but variable.
Standard Error
How much sample means vary from sample to sample. As n increases, the standard error decreases — larger samples are more precise.
The Big Picture
The backbone of statistical reasoning:
- We sample because we can't study everything
- Random sampling ensures representativeness
- We use statistics to estimate parameters
- The standard error measures our confidence
What you learned
Populations are the whole; samples are the part we study. Random sampling minimizes bias. Parameters describe populations; statistics describe samples. The standard error quantifies precision.