In statistics, a population is the entire group you want to learn something about. That group doesn’t have to be people. It can be a collection of measurements, events, objects, or outcomes, as long as it’s clearly defined. If you’re studying the average height of adults in the United States, every adult in the country is your population. If you’re studying how long lightbulbs last before burning out, every lightbulb of that type is your population. The word “population” in statistics is broader and more flexible than its everyday meaning.
How It Differs From the Everyday Meaning
When most people hear “population,” they think of a count of people living in a country or city. In statistics, the term refers to any complete set of items you’re interested in studying. The BMJ puts it plainly: statisticians speak of a population of objects, events, procedures, or observations. That includes things like blood pressure readings from every patient in a hospital, every transaction on an e-commerce site in a given month, or every surgical operation performed at a clinic over five years.
What makes something a statistical population is not its size or what it’s made of. It’s the fact that you’ve defined it as the full set of cases relevant to your question. A bag of 10 labeled chips can be a population. So can every voter enrolled in a country at the time of an election. The key requirement is that the population has clear boundaries: explicit rules about what’s included and what isn’t.
Finite vs. Infinite Populations
Some populations are countable. Every student enrolled at a university this semester, every registered voter in a state, every chip in a bag. These are finite populations. You could, in theory, list every single member.
Other populations are harder to pin down. If a psychologist wants to study how human memory works, the population of interest isn’t just people alive today. It could include any human being, past, present, or future. That’s an infinite (or at least indefinite) population. You can never observe all its members, because the group keeps growing or is defined so broadly that its edges are theoretical. Most research in medicine and psychology deals with populations like this, which is why sampling becomes essential.
Population vs. Sample
Studying an entire population directly is usually impractical. If you wanted to measure the blood pressure of every adult in a country, you’d need enormous time and resources. Instead, researchers select a smaller group called a sample, measure that group, and use the results to draw conclusions about the larger population.
This process is called statistical inference. The logic works like this: if your sample is selected properly (usually through random selection), the patterns you observe in the sample should closely reflect what’s true for the population as a whole. A well-chosen sample of 1,000 voters can give you a reliable estimate of how millions of people plan to vote, within a known margin of error.
The distinction between population and sample also shows up in notation. Values that describe a population are called parameters and use Greek letters. The population mean is written as μ (mu) and the population proportion as p. Values calculated from a sample are called statistics and typically use Latin letters, like x̄ for the sample mean. This notation exists so that readers always know whether a number describes the whole group or just the subset that was measured.
Target Population vs. Study Population
In practice, researchers often work with two versions of a population. The target population is the full group they want their findings to apply to. If you’re studying a new treatment for stroke, your target population might be every person worldwide who has had that type of stroke. But you can’t reach all of them, so you narrow it down to an accessible group: patients at specific hospitals during a specific time period, for instance. This smaller, reachable group is sometimes called the study population.
The gap between the target population and the study population matters. If your study population differs from your target population in important ways (younger, healthier, more urban), your results may not generalize as broadly as you’d like. This is why researchers define their populations carefully, with explicit inclusion and exclusion criteria. Those criteria spell out exactly who qualifies: age ranges, diagnoses, geographic location, and anything else that draws a boundary around the group being studied.
The Sampling Frame
Between the population and the sample sits something called a sampling frame. This is the actual list or database from which participants are drawn. If your population is “all adults in a city,” your sampling frame might be the city’s phone directory or voter registration list. No sampling frame is perfect. Some members of the population will be missing from any real-world list, and some entries on the list might not belong to the population. The quality of your sampling frame directly affects how well your sample represents the population.
Think of it as three nested circles. The population is the largest: everyone or everything you care about. The sampling frame is a smaller circle inside it: the portion you can actually identify and reach. The sample is the smallest circle: the group you actually measure. Good research tries to make these three circles overlap as much as possible.
Why the Definition Matters
The way you define your population shapes every conclusion you can draw from your data. A vague or poorly defined population makes it unclear who your results apply to. If a study tests a new teaching method on college freshmen at one university, the population isn’t “all students everywhere.” It’s something closer to “freshmen at similar universities taking similar courses.” Stretching the findings beyond that requires additional evidence.
Population definition also affects fairness. If inclusion criteria unintentionally exclude certain demographic groups, the results may not apply to those groups. A drug tested only on men between 30 and 50 tells you little about how it works in women or older adults. Researchers are expected to examine their criteria for biases related to gender, race, income, and access to ensure the defined population isn’t narrower than it needs to be.
In short, “population” in statistics is not just a number of people. It’s the precise answer to the question: “Who or what, exactly, am I trying to learn about?” Every statistical method, from calculating an average to running a complex model, depends on that answer being clear.