1 Definition
1.1 Basic concept
In statistics, a mean is a single numerical summary intended to represent the central or typical value of a set of observations. It compresses a collection of data into one figure that can be used for description, comparison, and later analysis. The most familiar mean is the arithmetic mean, but the term also covers several related averages that are defined in different ways.
A mean is not merely a mechanical calculation. It reflects a choice about how to summarize data and which features of the data should matter most. For some datasets, a simple average is appropriate; for others, a different form of mean better captures the underlying pattern.
1.2 Notation and terminology
The arithmetic mean of a sample is commonly written as x̄, read as “x-bar.” The mean of a population is often written as μ, the Greek letter mu. In formulas, the arithmetic mean is generally expressed as the sum of all values divided by the number of values.
In ordinary language, “average” is often used as a synonym for mean, although in statistics the two are not always identical. The word “mean” may also be qualified by an adjective such as geometric, harmonic, or quadratic, indicating the method used to combine the values.
1.3 Mean as a measure of central tendency
The mean belongs to the family of measures of central tendency, which aim to describe the center of a distribution. Compared with other measures in this family, the mean uses every observation in the dataset, not just one or two representative points. This makes it sensitive to the full structure of the data.
Because it is based on all observations, the mean is often useful when values are fairly evenly distributed. It can be less representative when the data are strongly asymmetric or contain extreme values.
2 Arithmetic mean
2.1 Formula
The arithmetic mean is computed by adding the values in a dataset and dividing the total by the number of values. If the values are denoted by x1, x2, ..., xn, the sample mean is
x̄ = (x1 + x2 + ... + xn) / n
For a population, the corresponding notation is usually
μ = (sum of all values) / N
where N is the population size.
2.2 Calculation for ungrouped data
For ungrouped data, each observation is listed individually. To find the arithmetic mean, one adds all values and divides by how many values there are. For example, the mean of 4, 6, and 10 is 20 divided by 3, which equals 6.67.
This direct method is straightforward and is the standard approach when the dataset is not large. It also shows that every data point contributes equally to the final result.
2.3 Calculation for grouped data
Grouped data are organized into classes or intervals rather than individual values. In this case, the mean is usually approximated by using the midpoint of each class and weighting it by the class frequency. The result is an estimate rather than an exact mean, because the precise values within each group are unknown.
This method is widely used in frequency tables and summarized datasets. It allows analysts to obtain a useful central value even when the raw observations are not available.
2.4 Weighted mean
2.4.1 Definition
A weighted mean is a mean in which some values contribute more strongly than others. Instead of treating every observation equally, each value is assigned a weight that indicates its relative importance, frequency, or reliability.
2.4.2 Formula
The weighted mean is calculated by multiplying each value by its weight, adding the products, and dividing by the sum of the weights:
weighted mean = [Σ(wi xi)] / [Σwi]
where xi are the values and wi are the weights.
2.4.3 Applications
Weighted means are used when observations represent different amounts of information. Common examples include grade-point averages, survey data with sampling weights, and financial indices. They are also useful when combining subgroup means that are based on different sample sizes.
3 Other types of means
3.1 Geometric mean
3.1.1 Definition
The geometric mean is a multiplicative average rather than an additive one. It is especially appropriate for quantities that grow by ratios or percentages.
3.1.2 Formula
For positive numbers x1, x2, ..., xn, the geometric mean is
(x1 x2 ... xn)^(1/n)
It is defined only for nonnegative or, more precisely, positive values in its standard statistical use.
3.1.3 Uses
The geometric mean is commonly used in finance, growth-rate analysis, and multiplicative processes. It is helpful when describing compounded change, such as long-term investment returns or population growth over several periods.
3.2 Harmonic mean
3.2.1 Definition
The harmonic mean is based on reciprocals and is especially suited to averaging rates. It gives greater influence to smaller values than the arithmetic mean does.
3.2.2 Formula
For positive numbers x1, x2, ..., xn, the harmonic mean is
n / [(1/x1) + (1/x2) + ... + (1/xn)]
3.2.3 Uses
The harmonic mean is often used for speeds, densities, and other rate-like quantities. For instance, it can summarize average speed over equal distances more accurately than the arithmetic mean.
3.3 Quadratic mean
3.3.1 Definition
The quadratic mean, also called the root mean square, is computed by squaring the values, averaging the squares, and taking the square root. It is especially relevant when larger deviations need greater emphasis.
3.3.2 Relationship to other means
For positive data, the quadratic mean is generally at least as large as the arithmetic mean, which is in turn at least as large as the harmonic mean. This ordering reflects the different ways each mean weights larger and smaller values. The quadratic mean is widely used in physics, engineering, and error analysis.
4 Properties
4.1 Sensitivity to extreme values
The arithmetic mean is affected by unusually large or small observations. A single extreme value can shift it noticeably, especially in a small dataset. This sensitivity is useful when all values should matter equally, but it can also make the mean less representative of the main body of the data.
4.2 Effect of adding or removing data
Adding a new observation changes the mean depending on how far that observation lies from the current mean. A value above the existing mean increases it, while a value below the existing mean decreases it. Removing a value has the opposite effect.
The mean of combined groups can be found from the group means and their sizes. This makes the mean convenient for updating summaries when data are accumulated in stages.
4.3 Algebraic properties
The arithmetic mean has several useful algebraic features. It behaves linearly under addition and scaling, which makes it easy to manipulate in formulas and statistical derivations. It is also the value that minimizes the sum of squared deviations from the data, a property that underlies many methods in estimation and regression.
5 Interpretation and use
5.1 Descriptive statistics
In descriptive statistics, the mean provides a compact summary of the center of a dataset. It is often reported alongside measures of spread such as the range, variance, or standard deviation. Together, these summaries help describe both the typical value and the variability of the data.
5.2 Comparison across samples
Means are frequently used to compare groups. For example, researchers may compare average test scores, heights, or production outputs across samples. Such comparisons are useful, but they should be interpreted in light of sample size, variability, and the shape of the distribution.
5.3 Role in inferential statistics
The mean is central to many inferential procedures. Sample means are used to estimate population means, and differences between means often form the basis of hypothesis tests and confidence intervals. Because of this role, the mean is one of the most important summary measures in statistical modeling.
6 Limitations
6.1 Influence of outliers
A major limitation of the mean is its susceptibility to outliers. Extreme values can distort the result and make it less typical of the majority of observations. In such cases, other summaries may better represent the center of the data.
6.2 Use with skewed distributions
When a distribution is highly skewed, the mean may lie away from the most common values. It can be pulled toward the long tail of the distribution, producing a center that does not match what many people would regard as “typical.” This is one reason why analysts often report the median as well.
6.3 Inappropriate data types
Means are not suitable for all kinds of data. They are meaningful for numerical quantities measured on interval or ratio scales, but not for nominal categories. For example, it does not make sense to compute a mean of colors, names, or other purely categorical labels.
7 Related concepts
7.1 Median
The median is the middle value in an ordered dataset. Unlike the mean, it is resistant to extreme observations and is often preferred for skewed distributions.
7.2 Mode
The mode is the most frequently occurring value. It is useful for identifying the most common category or the most common numerical value in a dataset.
7.3 Midrange
The midrange is the average of the smallest and largest values in a dataset. It is easy to compute but depends entirely on the extremes, making it less stable than the mean.
7.4 Expected value
Expected value is the probabilistic counterpart of the mean. In probability theory, it represents the long-run average outcome of a random variable under a specified distribution.