# Parameter Definition in Statistics: Meaning and Examples

The parameter definition in statistics is simple: a parameter is a numerical summary of an entire population. It describes a fixed, usually unknown value such as a population mean or standard deviation. Because measuring every member of a population is often impossible, analysts estimate parameters from samples.

## Quick Answer

- A parameter is a number that describes a whole population, for example the population mean $\mu$ or population standard deviation $\sigma$ [1].
- A statistic is a number that describes a sample, for example the sample mean $\bar{x}$ or sample standard deviation $s$ [2].
- Parameters are fixed but usually unknown, so you estimate them from sample data [1].
- Parameters are written with Greek letters ($\mu$, $\sigma$, $p$) and statistics with Latin letters ($\bar{x}$, $s$, $\hat{p}$).
- The accuracy of an estimate depends on sample size and how the sample was drawn.

## What a Parameter Means

In everyday language, a parameter is a limit or a boundary. In statistics, the meaning is narrower and more precise. A parameter is any quantity of a statistical population that summarizes or describes an aspect of that population, such as a mean or a standard deviation [1].

The precise statistical definition: a parameter is a numerical characteristic of a population of size $N$ that can only be obtained by a census, and it stays constant although it is usually unknown [2]. If a population follows a known distribution, such as the normal distribution, a small set of parameters fully describes it and defines the probability distribution used to draw samples [1].

Two ideas matter here. First, a parameter belongs to the population, not to any sample. Second, its true value is fixed. It does not change when you collect new data. What changes is your estimate of it.

## How It Works

The mechanism is estimation. You cannot usually measure the whole population, so you compute a statistic from a sample and use it as an estimate of the matching parameter [1]. The table below shows the common pairings.

| Population parameter | Symbol | Sample statistic | Symbol |
|---|---|---|---|
| Mean | $\mu$ | Sample mean | $\bar{x}$ |
| Standard deviation | $\sigma$ | Sample standard deviation | $s$ |
| Proportion | $p$ | Sample proportion | $\hat{p}$ |
| Variance | $\sigma^2$ | Sample variance | $s^2$ |

The population mean is defined as:

$$\mu = \frac{\sum x}{N}$$

Each symbol means the following. $\sum x$ is the sum of every value in the population. $N$ is the population size. $\mu$ is the parameter you want.

The population standard deviation is:

$$\sigma = \sqrt{\frac{\sum (x - \mu)^2}{N}}$$

Here $x - \mu$ is each value's deviation from the population mean, and dividing by $N$ averages the squared deviations before the square root. The sample standard deviation uses $n - 1$ in the denominator instead of $n$, which corrects for the fact that a sample tends to be less spread out than its population.

Parameters also appear in models. In a straight-line regression with an unknown intercept and slope, there are two parameters and one predictor variable, written $f(x;\vec{\beta}) = \beta_0 + \beta_1 x$ [3]. The slope and intercept are the parameters defining the model, and they have no meaning for individuals even though they can predict an individual's value [4].

## Worked Example

The dataset is a survey of scores from 0 to 100 for a class of 20 students. The full class is the population.

| student_id | score | student_id | score |
|---|---|---|---|
| 1 | 72 | 11 | 62 |
| 2 | 65 | 12 | 81 |
| 3 | 88 | 13 | 77 |
| 4 | 91 | 14 | 70 |
| 5 | 55 | 15 | 86 |
| 6 | 78 | 16 | 59 |
| 7 | 83 | 17 | 93 |
| 8 | 69 | 18 | 68 |
| 9 | 74 | 19 | 80 |
| 10 | 95 | 20 | 84 |

Step 1. Population size: $N = 20$.

Step 2. Sum of all scores: $\sum x = 1530$.

Step 3. Population mean parameter: $\mu = 1530 / 20 = 76.5000$.

Step 4. Population standard deviation parameter: $\sigma = 11.1557$.

Step 5. Draw five samples of five students each and compute the sample mean statistic for each.

| Sample | Sample mean $\bar{x}$ | Sample SD $s$ |
|---|---|---|
| 1 | 77.8000 | 8.5264 |
| 2 | 81.6000 | 11.1041 |
| 3 | 72.0000 | 12.8647 |
| 4 | 74.6000 | 14.0107 |
| 5 | 75.8000 | 10.5214 |

The mean of the five sample means is 76.36, close to the true parameter of 76.5000 but not equal to it. That gap is sampling error.

The same values come out of a spreadsheet. `=AVERAGE(A2:A21)` returns 76.5000, `=STDEV.P(A2:A21)` returns 11.1557 for the population parameter, and `=STDEV.S(A2:A21)` returns 11.4455 for the sample statistic.

```python
import statistics
pop = [72, 65, 88, 91, 55, 78, 83, 69, 74, 95, 62, 81, 77, 70, 86, 59, 93, 68, 80, 84]
mu = statistics.mean(pop)          # parameter
s  = statistics.stdev(pop)         # sample SD (n-1)
sample = [72, 91, 69, 77, 80]    # sample 1: students 1, 4, 8, 13, 19
xbar = statistics.mean(sample)     # statistic
print(f"{mu:.4f} {xbar:.4f}")  # 76.5000 77.8000
```

Output:

```
μ = 76.5000 (parameter), x̄₁ = 77.8000 (statistic); σ = 11.1557, s = 11.4455
```

## How to Interpret It

Read a parameter as the truth you are trying to reach. The value 76.5000 is the real class mean because every student was measured. The value 77.8000 is only what one sample suggests. If you had drawn a different five students, you would have gotten a different number, as the range from 72.0000 to 81.6000 shows.

Two habits help. First, always ask whether a number describes a population or a sample, because that decides whether it is a parameter or a statistic. Second, treat sample estimates as uncertain. A single sample mean is a point estimate, and its distance from the parameter shrinks as the sample grows. This is the core idea behind [descriptive vs inferential statistics](/blog/data-analysis/descriptive-vs-inferential-statistics), where sample summaries are used to reason about populations.

## When to Use It (and when not to)

Use parameters when you need to describe a whole group precisely, when you are building a probability model, or when you want to state what a sample is estimating. If a population exactly follows a known distribution, a small set of parameters gives a complete description of it [1].

Do not treat a parameter as something you can always observe. Its true value is unknown and unknowable except in simulation experiments [3]. Also do not confuse a parameter with a variable. A variable is a measured characteristic that differs between individuals, while a parameter is a fixed summary of the population [4]. If you only care about the sample in front of you, a statistic is enough and no parameter is needed.

## Parameter vs Statistic

The closest related idea is the statistic. Both are numerical summaries, but they describe different things.

| Feature | Parameter | Statistic |
|---|---|---|
| Describes | Population | Sample |
| Symbol style | Greek ($\mu$, $\sigma$, $p$) | Latin ($\bar{x}$, $s$, $\hat{p}$) |
| Value | Fixed, usually unknown | Varies from sample to sample |
| How obtained | Census | Sample data |
| Role | Target of estimation | Estimate of the target |

A parameter is the fixed target. A statistic is the moving arrow you shoot at it. For a deeper look at one of these statistics, see [sample mean](/blog/data-analysis/sample-mean).

## Common Mistakes

- Calling a sample mean a parameter. If the number came from part of the group, it is a statistic. Fix: check whether every member was measured.
- Assuming the parameter changes when new data arrive. The parameter is fixed, your estimate moves. Fix: separate the true value from the estimate in your wording.
- Using $n$ instead of $n - 1$ for a sample standard deviation. This biases the estimate downward. Fix: use the sample formula for samples and the population formula only for full populations.
- Treating a point estimate as exact. A sample mean of 77.8000 is not the population mean. Fix: report an interval or the sample size alongside it.
- Mixing up a parameter with a variable. A variable varies across individuals, a parameter does not [4]. Fix: ask whether the quantity is a property of the group or of a person.
- Ignoring how the sample was drawn. A biased sample gives a biased estimate no matter how large it is. Fix: check the sampling method before trusting the number.

## Limitations

A parameter cannot usually be known with certainty. In most real settings it is unobservable, and the analyst can only estimate or infer it from a random sample [1]. Even a well-drawn sample gives an estimate that differs from the truth by some amount, and that amount is not visible from the sample alone.

Parameters also depend on the model you assume. If you assume a normal distribution and the data are not normal, the parameters you estimate describe the wrong family of distributions. Nonparametric methods avoid assuming a particular family, but then they do not estimate parameters for such a distribution [4]. Finally, a parameter summarizes a population, so it says nothing about any single individual.

## Frequently Asked Questions

### What is a parameter in simple terms?

A parameter is a number that describes a whole population, like the average score of every student in a class. It is fixed but often unknown, so you estimate it from a sample. The matching number from the sample is called a statistic.

### What is the difference between a parameter and a statistic?

A parameter describes the population and a statistic describes a sample [2]. Parameters are written with Greek letters and statistics with Latin letters. The statistic is used to estimate the parameter.

### Is the population mean a parameter?

Yes. The population mean $\mu$ is a parameter because it summarizes every value in the population. The sample mean $\bar{x}$ is the statistic you use to estimate it.

### Why are parameters usually unknown?

Measuring every member of a population is often impractical or impossible, so the true value stays hidden [1]. You work with samples and accept some uncertainty. Only a full census or a simulation reveals the true parameter.

### What does parametric mean?

Parametric methods estimate parameters of an underlying theoretical distribution, which is why t tests are called parametric [4]. Nonparametric methods do not assume a particular distribution family and so do not estimate parameters for one.

## References

1. [Statistical parameter - Wikipedia](https://en.wikipedia.org/wiki/Statistical_parameter)
2. [3.2: Parameter and Statistic - Mathematics LibreTexts](https://math.libretexts.org/Courses/Mt._San_Jacinto_College/Interactive_Lecture_Notes_for_Introductory_Statistics/03%3A_Numerical_Summaries_of_Data/3.02%3A_Parameter_and_Statistic)
3. [4.1.2. What terminology do statisticians use to describe process models?](https://www.itl.nist.gov/div898/handbook/pmd/section1/pmd12.htm)
4. [Statistics notes Variables and parameters - PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC1116021/)

## Further Reading

- [NIST/SEMATECH e-Handbook of Statistical Methods](https://www.itl.nist.gov/div898/handbook/index.htm)
- [OpenStax. Introductory Statistics 2e](https://openstax.org/details/books/introductory-statistics-2e)

## Related Articles

- [What Is the Mode in Statistics? Definition and Examples](/blog/data-analysis/what-is-mode-statistics-definition)
- [What Is Descriptive Statistics? Definition and Examples](/blog/data-analysis/descriptive-statistics)
- [What Is an Ordinal Scale? Definition and Examples](/blog/data-analysis/ordinal-scale-definition-examples)
- [Sample Mean: Definition, Formula and Examples](/blog/data-analysis/sample-mean)
- [What Is a Stratum? Definition and Examples in Statistics](/blog/data-analysis/stratum-definition-statistics-examples)
- [Statistical Parameter: Definition, Types, and Estimation](/blog/guides/statistical-parameter-definition-types-and-estimation)
- [What Does Statistical Mean? A Clear Explanation](/blog/guides/what-does-statistical-mean-a-clear-explanation)
- [Statistical Synonyms: A Guide to Terminology in Statistics](/blog/guides/statistical-synonyms-a-guide-to-terminology-in-statistics)