Microsoft Office Tutorials and References

In Depth Information

**Using NORM. DIST and POISSON. DIST to Determine Probabilities**

POISSON.DIST

Poisson is another kind of distribution used in many areas of statistics. Its

most common use is to model the number of events taking place in a

specified time period. Suppose you were modeling the number of employees

calling in sick each day, or the number of defective items produced at your

factory each week. In these cases, the Poisson distribution is appropriate.

The
Poisson
distribution is useful for analyzing rare events. What exactly

does
rare
mean? People calling in sick at work is hardly rare, but a specific

number calling in sick is rare, at least statistically speaking. Situations where

Poisson is applicable include numbers of car accidents, counts of customers

arriving, manufacturing defects, and the like. One way to express it is that the

events are individually rare, but there are many opportunities for them to

happen.

The Poisson distribution is a discrete distribution. This means that the X

values in the distribution can only take on specified, discrete values such as

X = 1, 2, 3, 4, 5 and so on. This is different from the normal distribution, which

is a continuous distribution in which X values can take any value (X = 0.034,

1.2365, and so on). The discrete nature of the Poisson distribution is suited to

the kinds of data you use it with. For example, with employees calling in sick,

you may have 1, 5, or 8 on a given day, but certainly not 1.45, 7.2, or 9.15!

Figure 11-9 shows a Poisson distribution that has a mean of 20. Values on the

X axis are number of occurrences (of whatever youâ€™re studying), and values

on the Y axis are probabilities. You can use this distribution to determine

the probability of a specific number of occurrences happening. For example,

this chart tells you that the probability of having exactly 15 occurrences is

approximately 0.05 (5 percent).

Figure 11-9:

A Poisson

distribution

with a

mean of 20.