Population vs Sample: What’s the Difference?

A population is every person or thing you want to draw conclusions about. A sample is the part of it you actually measure. Nearly all statistics is about using a sample to say something about a population you can’t measure in full.

Examples

QuestionPopulationSample
How satisfied are our customers?Everyone who bought from us this yearThe 400 who answered the survey
Does the new drug lower blood pressure?All adults with high blood pressureThe 120 people in the trial
What share of bolts from this machine are faulty?Every bolt the machine makesThe 50 bolts pulled from today’s batch
What is the average grade in my class of 28?The 28 studentsNone needed: you can measure all 28

The last row matters. If you have measured every member of the group you care about, you have the population, and there is nothing to infer. You can just describe it.

Why the formulas change: n versus n − 1

The difference between population and sample shows up most clearly in the standard deviation. Take the values 4, 8, 6, 5 and 7. The mean is 6 and the squared distances from the mean add up to 10.

  • If these five values are the whole population, divide by n: variance = 10 ÷ 5 = 2, and the standard deviation is 1.414.
  • If they are a sample from a bigger group, divide by n − 1: variance = 10 ÷ 4 = 2.5, and the standard deviation is 1.581.

See both versions side by side in the standard deviation calculator

Why n − 1? A sample’s values sit closer to their own mean than to the true population mean, because the sample mean was calculated from those very values. Dividing by n would give a figure that is too small on average. Dividing by n − 1 corrects for that. With large samples the difference is tiny. With five values it is about 12%.

Excel has both versions: STDEV.S for a sample and STDEV.P for a population. If you are unsure which to use, you almost certainly want the sample version.

Notation

Population (parameter)Sample (statistic)
SizeNn
Meanμx̄
Standard deviationσs
Proportionpp̂
Correlationρr

Greek letters for the population, Latin letters for the sample. See parameter vs statistic for more on this.

What makes a good sample

A large sample doesn’t fix a biased one. A famous example: in 1936 the Literary Digest polled over two million people and predicted Alf Landon would beat Franklin Roosevelt. Roosevelt won in a landslide. The magazine had drawn its list from car owners and telephone directories, which in the 1930s leaned towards wealthier voters.

What matters most is how the sample was picked. In a simple random sample, every member of the population has the same chance of being chosen. Surveys that people opt into, such as website polls, tend to attract those with strong opinions.

How big should a sample be? That depends on how precise you need to be. The sample size guide walks through it.

Check yourself

A teacher records the test scores of all 25 students in her class to report the class average. Population or sample?

Population, as long as the class is all she wants to talk about. If she wants to say something about all students taking this test nationally, her class is a sample, and not a random one.

Why can’t a sample of one give you a sample standard deviation?

With n = 1, n − 1 = 0, and you can’t divide by zero. One value tells you nothing about spread.