Patterns of Distribution
Much of the data we use are points along a measurement scale that have values. These points can range from lower values to higher values. The patterns of distribution of values from cases, or data points, can form one of four patterns.
NORMAL DISTRIBUTION: The most common pattern is the normal or bell-shaped curve. As a
general description this distribution pattern looks like a pile formed when salt
is poured from a central point and the salt makes a pile. This pattern is the result of the binomial distribution producing a
standard-normal distribution based on probability. Most people know the
simplest level of this distribution as that from a coin toss. If a coin is
tossed twice times, or two coins tossed once, probability theory predicts a 25%
chance of two heads or two tails, and 50% chance of one head and one tail.
Continuing this pattern out produces tables of probabilities for such normal
distributions. 
UNIFORM DISTRIBUTION: Another common pattern appears when all of the cases are
spread out along a series of values each of which has the same number of cases.
A descriptions of this patter looks like sand spread out in a flat bowl. This patterns is the result of each value having equal probability. Most
people know the simplest level of this distribution as that from a dice toss.
The probability of each of the values from 1 to 6 is equally likely when a
single die is tossed. 
SKEWED
DISTRIBUTIONS: Two other common patterns appear when a normal
distribution is not symmetrical. That is, one side (called a tail) is
extended out and the other side is drawn in, thus the high point is not centered
among the cases. This pattern looks like a pile blown by the wind, like
drifting sand dunes.When the tail drifts off to the higher values it is called a positive
skew. The red image to the left is an illustration of a positive skew.
As you would expect, when the tail drifts off to the lower values it is called a negative skew.
BI-MODAL DISTRIBUTIONS: When a pattern has two high points, it is called a bi-modal distribution (or multi-modal if there are more than two peaks). In this case the peaks the best measure of an average is not a central point but a list of the modes.
Among the most commonly used measures of dispersion of cases are the related concepts of standard deviation and variance.
