What the Mean of a Probability Distribution Represents
The mean of a probability distribution is the long-run average value you expect if you could repeat an experiment infinitely many times. It summarizes the central tendency of a random variable by weighting each possible outcome by its probability. Unlike a sample mean from observed data, the distribution mean is a theoretical parameter that describes the center of the underlying probabilistic model. This concept applies to both discrete and continuous random variables, providing a single number that represents the balance point of the distribution.
Definitions and Core Concepts
Probability Distribution and Expected Value
A probability distribution assigns probabilities to all possible outcomes of a random experiment. The mean, formally called the expected value, is computed as a weighted average where each outcome is multiplied by its probability. For a discrete random variable, this involves summing over all possible values; for a continuous random variable, it involves integration across the range of values. The mean is a measure of location and one of the key descriptive characteristics used alongside variance and standard deviation to understand the behavior of a distribution.
Notation and Key Terms
- Random variable: denoted by X, representing possible outcomes
- Probability mass function (PMF): for discrete variables, giving P(X = x)
- Probability density function (PDF): for continuous variables, describing relative likelihoods
- Expected value: often written as E[X] or μ (mu)
How to Calculate the Mean for Discrete Distributions
To find the mean of a discrete probability distribution, multiply each possible value by its probability and sum the results. This is expressed mathematically as E[X] = Σ [x · P(X = x)], where the sum is taken over all possible values x. The calculation requires that you know or can derive the probability for each outcome, and that the probabilities sum to one. This formula is foundational and applies to distributions such as the binomial, Poisson, and multinomial.
Worked Example and Table
Consider a fair six-sided die where X is the number shown. Each face has probability 1/6. The mean is (1 + 2 + 3 + 4 + 5 + 6) / 6 = 3.5. Even though 3.5 is not an observable outcome, it represents the average over many rolls.
| Outcome (x) | P(X = x) | x · P(X = x) |
|---|---|---|
| 1 | 1/6 | 1/6 |
| 2 | 1/6 | 2/6 |
| 3 | 1/6 | 3/6 |
| 4 | 1/6 | 4/6 |
| 5 | 1/6 | 5/6 |
| 6 | 1/6 | 6/6 |
Sum of x · P(X = x) = 21/6 = 3.5, which is the mean of the distribution.
How to Calculate the Mean for Continuous Distributions
For continuous random variables, the mean is found by integrating the product of the variable and its probability density function (PDF) over the entire range. The formula is E[X] = ∫ x · f(x) dx, where f(x) is the PDF and the integral is taken over all real numbers for which f(x) > 0. Common continuous distributions include the normal, exponential, and uniform, each with a known mean derived from their parameters. Integration replaces summation to account for uncountably many possible values.
Example: Exponential Distribution
The exponential distribution with rate parameter λ has PDF f(x) = λ e^{-λ x} for x ≥ 0. Its mean is E[X] = 1/λ. This result comes from evaluating the integral ∫ x · λ e^{-λ x} dx from 0 to infinity using integration by parts. The mean provides the average waiting time between events in Poisson processes, such as time between customer arrivals.
Properties and Interpretation of the Distribution Mean
Linearity and Transformations
The expected value operator is linear, meaning E[aX + bY] = aE[X] + bE[Y] for random variables X and Y and constants a and b. This property holds regardless of dependence between X and Y, though variance calculations must account for covariance. For transformations, E[g(X)] equals Σ g(x) P(X = x) (discrete) or ∫ g(x) f(x) dx (continuous), which is not generally the same as g(E[X]) unless g is linear.
Relationship to Other Statistics
The distribution mean is one measure of center. For symmetric distributions, the mean coincides with the median and mode, but for skewed distributions it can differ substantially. The mean is sensitive to extreme values, unlike some robust estimators, which makes it powerful for modeling but potentially influenced by outliers. It also serves as the center of balance for the distribution, the point at which the weighted deviations sum to zero.
Practical Applications and Common Distributions
In practice, you often work with named distributions whose means are known from theory. For a binomial distribution with n trials and success probability p, the mean is n · p. For a normal distribution with parameters μ and σ, the mean is simply μ. For a uniform distribution on [a, b], the mean is (a + b) / 2. These standard results let you compute means quickly without performing calculus each time, provided the problem matches the distributional assumptions.
Comparison of Means for Common Distributions
| Distribution | Mean | Parameters |
|---|---|---|
| Bernoulli | p | Success probability p |
| Binomial | n · p | Number of trials n, success probability p |
| Poisson | λ | Rate parameter λ |
| Uniform (continuous) | (a + b) / 2 | Lower bound a, upper bound b |
| Normal | μ | Location μ, scale σ |
| Exponential | 1 / λ | Rate parameter λ |
Step-by-Step Guide to Finding the Mean
- Identify whether the random variable is discrete or continuous.
- Obtain the probability mass function (PMF) or probability density function (PDF).
- Verify that the probabilities or density integrate to one.
- Set up the sum (discrete) or integral (continuous) using E[X] = Σ x · P(X = x) or E[X] = ∫ x · f(x) dx.
- Compute the sum or integral using algebraic techniques or known formulas.
- Interpret the mean as the long-run average over many repetitions.
Common Pitfalls and Considerations
One frequent mistake is confusing the sample mean with the distribution mean. The sample mean is computed from observed data, while the distribution mean is a theoretical parameter. Another issue is applying the mean formula outside the distribution’s support, such as using 1/λ for an exponential distribution defined only for positive x. Ensure the PDF or PMF is properly normalized before calculating the mean. Also note that the mean may not exist or be finite for heavy-tailed distributions, where integration or summation diverges.
When the Mean Alone Is Not Enough
While the mean summarizes central tendency, it does not describe spread or shape. Two distributions can share the same mean but differ greatly in variability or skewness. Use the variance, standard deviation, and higher moments to capture dispersion and asymmetry. Visual tools such as histograms and density plots complement numerical summaries and help assess whether the mean is a representative summary of the distribution.
Frequently Asked Questions
- Can the mean of a distribution be outside the range of possible values? Yes, for example the mean of a Bernoulli random variable is p, which can lie between 0 and 1 even though outcomes are only 0 or 1.
- Does the mean always equal a probable outcome? Not necessarily; the mean is a weighted average and need not correspond to any actual observable value.
- How does sample size affect the sample mean versus the distribution mean? As sample size grows, the sample mean converges to the distribution mean by the law of large numbers.
Summary
Finding the mean of a probability distribution involves computing a weighted average of all possible values, with probabilities serving as weights. For discrete variables this is a sum, and for continuous variables this is an integral. Named distributions have standard mean formulas that simplify calculations. Understanding the mean provides a foundation for further statistical analysis, but it should be considered alongside measures of variability and the overall shape of the distribution.