Moments in Statistics

Definition

In statistics, a Moment is a quantitative measure of the shape of a set of points (a distribution). Moments describe various characteristics of a frequency distribution, such as its central location (mean), dispersion (spread), asymmetry (skewness), and "peakedness" (kurtosis). Mathematically, the $r$-th moment of a variable $X$ about a constant $a$ is defined as the expected value of $(X - a)^r$.

Main Content

1. Raw Moments (Moments about the Origin)

Raw moments are calculated by taking the moments about zero ($a=0$).
They are denoted by $\mu'_r$ and are defined as $E(X^r)$. The first raw moment ($\mu'_1$) is the arithmetic mean.

2. Central Moments (Moments about the Mean)

Central moments are calculated by taking the moments about the mean of the data ($\bar{X}$ or $\mu$).
They are denoted by $\mu_r$ and are defined as $E[(X - \mu)^r]$. The second central moment ($\mu_2$) represents the variance of the distribution.

3. Standardized Moments

These are dimensionless quantities obtained by dividing the central moments by the standard deviation raised to the power of $r$.
They are used to compare the shapes of distributions regardless of the scale of measurement. For example, Skewness is the 3rd standardized moment, and Kurtosis is the 4th standardized moment.

Visual representation of data distribution around the mean (μ):
       |
       |     _---_
       |    /     \
       |   /       \
-------|--+---------+----
       μ-σ  μ     μ+σ
(Spread based on 2nd moment)

Working / Process

1. Calculate the Arithmetic Mean

First, sum all the observations in the dataset.
Divide the total sum by the number of observations ($n$) to find the mean ($\bar{x}$).

2. Calculate Deviations

For Central Moments, subtract the mean found in Step 1 from every individual data point ($x_i - \bar{x}$).
For Raw Moments, simply use the original data points without subtraction.

3. Compute Power and Average

Raise these deviations (or raw values) to the $r$-th power (e.g., square them for $r=2$, cube them for $r=3$).
Calculate the average of these powered values to obtain the specific moment.

Advantages / Applications

Descriptive Analytics: Moments provide a concise summary of complex datasets, identifying the center and the spread.
Skewness Determination: The third moment ($r=3$) identifies whether a distribution is symmetric or tilted toward one side.
Kurtosis Assessment: The fourth moment ($r=4$) measures the "tailedness" of the distribution, helping identify outliers.
Probability Theory: Moments are used to define Moment Generating Functions (MGFs), which are essential for identifying the probability distribution of random variables.

Summary

Moments are mathematical descriptors used to define the fundamental characteristics of a probability distribution. By calculating raw moments about zero and central moments about the mean, statisticians can quantify the center, variance, symmetry, and peak intensity of data. Key terms to remember: Raw Moments, Central Moments, Standardized Moments, Skewness, and Kurtosis.