Welcome to the Poisson Distribution!

Hello! Today we are diving into one of the most useful tools in statistics: the Poisson Distribution. While the name might sound a bit fancy (it’s French for "fish" and pronounced "pwa-son"), it is actually a very friendly and straightforward way to count things.

In this chapter, we are going to learn how to predict how many times an event happens within a fixed amount of time or space. Think of things like how many emails you get in an hour, or how many chocolate chips are in a cookie. Don't worry if this seems tricky at first; we will break it down piece by piece!

1. When Can We Use Poisson? (SB1)

The Poisson distribution is used to model the number of times a random event occurs in a fixed interval of time or space. However, we can only use it if the events meet four very specific conditions. You can remember these with the acronym "CRIS":

C – Constant Rate: The events must occur at a constant average rate (which we call \(\lambda\), the Greek letter "lambda").
R – Random: Events occur randomly and cannot be predicted individually.
I – Independent: One event happening doesn't make it more or less likely for another event to happen.
S – Singly: Events occur one at a time, never simultaneously (no "double-headers").

The Notation

When we want to say "The random variable \(X\) follows a Poisson distribution with a mean of \(\lambda\)," we write it like this:
\(X \sim \text{Po}(\lambda)\)

Quick Example

Imagine a quiet village road where, on average, 3 cars pass every hour. If the cars arrive independently and at a constant rate, we could say \(X \sim \text{Po}(3)\).

Quick Review: To use Poisson, events must be Independent, occur at a Constant rate, occur Singly, and be Random (CRIS).

2. The Poisson Formula (SB2)

If we know the average rate (\(\lambda\)), we can calculate the exact probability of a specific number of events (\(x\)) happening using this formula:

\[P(X = x) = \frac{e^{-\lambda} \lambda^x}{x!}\]

What do these symbols mean?
  • \(e\): This is a constant number (roughly 2.718). Your calculator has a special button for it!
  • \(\lambda\): The average number of events in the given interval.
  • \(x\): The number of successes we want to find the probability for (0, 1, 2...).
  • \(x!\): "x factorial" (e.g., \(4! = 4 \times 3 \times 2 \times 1\)).
Using Your Calculator

Most modern scientific and graphical calculators have a Poisson PD (Probability Density) and Poisson CD (Cumulative Distribution) function. In your exam, it is usually much faster to use these functions than the formula!

Did you know? The Poisson distribution was first used to model the number of soldiers in the Prussian army who were accidentally killed by horse kicks!

Key Takeaway: Use the formula for exact "equals" probabilities, or use your calculator’s Poisson functions to save time during the exam.

3. Mean, Variance, and Standard Deviation (SB3)

Here is some great news: The Poisson distribution is one of the easiest to work with when it comes to averages and spread!

For any Poisson distribution \(X \sim \text{Po}(\lambda)\):

Mean \(E(X) = \lambda\)
Variance \(\text{Var}(X) = \lambda\)

Because the variance is \(\lambda\), the Standard Deviation is simply \(\sqrt{\lambda}\).

Why is this important?

In your exam, if you are told that a distribution is Poisson and the mean is 5, you automatically know the variance is also 5. If a question shows you data where the mean and variance are very different, it’s a clue that the Poisson distribution might not be a good model!

Common Mistake: Students often forget to square root the variance to find the standard deviation. Remember: \(\text{SD} = \sqrt{\text{Variance}}\).

4. Adding Independent Poisson Distributions (SB4)

What if you are watching two different things at once? For example, the number of emails you get (\(X\)) and the number of texts you get (\(Y\))?

If \(X\) and \(Y\) are independent Poisson variables:

If \(X \sim \text{Po}(\lambda)\) and \(Y \sim \text{Po}(\mu)\), then:
\(X + Y \sim \text{Po}(\lambda + \mu)\)

Simply put: You just add the means together!

Real-World Example

If you get an average of 2 emails per hour and 3 texts per hour, the total number of notifications you get per hour follows a Poisson distribution with a mean of \(2 + 3 = 5\).

Key Takeaway: To combine independent Poisson events, just sum their average rates (\(\lambda\) values).

5. Hypothesis Testing with Poisson (SB5)

Sometimes we want to know if the average rate (\(\lambda\)) has changed. For example, has a new road safety campaign actually reduced the number of accidents?

Step-by-Step Hypothesis Test:

1. State the Hypotheses:
    \(H_0: \lambda = \text{old rate}\) (Nothing has changed)
    \(H_1: \lambda <, > \text{ or } \neq \text{old rate}\) (The rate has changed)
2. Assume \(H_0\) is true: Identify the distribution \(X \sim \text{Po}(\lambda)\).
3. Calculate the Probability: Find the probability of getting the observed value or more extreme.
    • If testing for a decrease: \(P(X \le \text{observed value})\)
    • If testing for an increase: \(P(X \ge \text{observed value})\)
4. Compare to the Significance Level: If your probability is less than the significance level (e.g., 0.05), the result is significant. 5. Write a Conclusion: "There is sufficient evidence to suggest that the rate of [context] has changed."

Important Tip for "Greater Than" Tests

When using a calculator or tables for \(P(X \ge k)\), remember that:
\(P(X \ge k) = 1 - P(X \le k - 1)\).
Example: The probability of getting 5 or more is 1 minus the probability of getting 4 or fewer.

Quick Review: Hypothesis testing checks if an observed result is "weird enough" to suggest the average rate has actually changed. Always include the context of the question in your final answer!