Welcome to Combinations of Random Variables!
In your previous Statistics units, you learned how to look at a single random variable, like the height of a student or the weight of an apple. But what happens if you put three apples in a bag? Or what if you want to find the difference between the weight of a box and the weight of the items inside?
In this chapter, we explore how to combine different independent random variables. This is a vital skill for Unit S3, as it forms the foundation for more advanced testing later on. Don't worry if it seems a bit abstract at first—once you see the patterns, it’s just like following a recipe!
1. Prerequisite: The Rules of the Game
Before we combine variables, let's quickly recap two rules from Statistics 1 (S1) that you'll need. If \(X\) is a random variable and \(a\) and \(b\) are constants:
- Expectation (The Mean): \(E(aX + b) = aE(X) + b\)
- Variance (The Spread): \(Var(aX + b) = a^2Var(X)\)
Quick Tip: Notice that for variance, the \(b\) disappears (adding a constant doesn't change the spread) and the \(a\) is squared because variance is measured in units squared!
2. Combining Two Different Variables
Imagine you have two independent variables: \(X\) (the weight of a coffee cup) and \(Y\) (the weight of the coffee inside). To find the total weight, we look at \(X + Y\).
The Mean of a Combination
Finding the new mean is very straightforward. You simply add or subtract them as you’d expect:
\(E(aX \pm bY) = aE(X) \pm bE(Y)\)
The Variance of a Combination
This is where students often trip up, so pay close attention! If \(X\) and \(Y\) are independent:
\(Var(aX \pm bY) = a^2Var(X) + b^2Var(Y)\)
Wait, why is there a plus sign for both?
Think of variance as "uncertainty" or "error." If you add two items, your uncertainty increases. If you subtract one item from another, your total uncertainty still increases because you are combining the "wobbliness" of both measurements.
Analogy: If you try to measure the gap between two shaky tables, the gap will be even shakier than the tables themselves!
Common Mistake to Avoid: Never subtract variances. Even if the formula asks for \(X - Y\), you add the variances: \(Var(X - Y) = Var(X) + Var(Y)\).
3. Linear Combinations of Normal Variables
The core of the S3 syllabus focuses on what happens when \(X\) and \(Y\) both follow a Normal Distribution.
The Golden Rule: If \(X\) and \(Y\) are Normal, then any linear combination of them (like \(X + Y\) or \(2X - 3Y\)) is also Normal.
The Master Formula
If \(X \sim N(\mu_x, \sigma_x^2)\) and \(Y \sim N(\mu_y, \sigma_y^2)\) are independent, then:
\(aX \pm bY \sim N(a\mu_x \pm b\mu_y, a^2\sigma_x^2 + b^2\sigma_y^2)\)
Step-by-Step Process for Problems:
- Identify the means (\(\mu\)) and variances (\(\sigma^2\)) for each variable.
- Calculate the new mean using the expectation rule.
- Calculate the new variance (remembering to square the coefficients and always add).
- Write down the new distribution in the form \(N(\text{new mean}, \text{new variance})\).
- Use your calculator or Normal distribution tables to find the required probability.
Example:
The weight of a cookie \(C \sim N(30, 2)\) and the weight of its packaging \(P \sim N(5, 0.5)\). What is the distribution of the total weight \(T = C + P\)?
\(E(T) = 30 + 5 = 35\)
\(Var(T) = 2 + 0.5 = 2.5\)
So, \(T \sim N(35, 2.5)\).
4. The "Sum" vs. "Multiple" Trap
This is one of the most common places to lose marks in Further Maths S3. There is a huge difference between one item multiplied by \(n\) and \(n\) separate items added together.
Case A: Multiple of one variable (\(nX\))
Imagine taking one giant 2kg bag of flour.
\(E(2X) = 2E(X)\)
\(Var(2X) = 2^2Var(X) = 4Var(X)\)
Case B: Sum of independent variables (\(X_1 + X_2\))
Imagine taking two separate 1kg bags of flour. Because they are separate, their variations might cancel each other out slightly (one might be slightly heavy, the other slightly light).
\(E(X_1 + X_2) = E(X) + E(X) = 2E(X)\)
\(Var(X_1 + X_2) = Var(X) + Var(X) = 2Var(X)\)
Key Takeaway: Adding independent items (Case B) results in a smaller variance than multiplying a single item (Case A). In exams, read carefully: are you buying "one 5-liter bottle" (\(5X\)) or "five 1-liter bottles" (\(X_1 + X_2 + X_3 + X_4 + X_5\))?
5. Summary and Quick Review
Did you know? This ability to combine normal variables is why we can use the Central Limit Theorem (which you will see later in S3) to make predictions about averages!
Quick Review Box:
- Means: \(E(aX + bY) = aE(X) + bE(Y)\)
- Variances: \(Var(aX + bY) = a^2Var(X) + b^2Var(Y)\)
- Variances: \(Var(aX - bY) = a^2Var(X) + b^2Var(Y)\) (Always ADD)
- Linear combinations of Normal variables are always Normal.
- Standard Deviation is the square root of Variance—always convert to Variance before using these formulas!
Encouragement: If the variance rules feel strange, just remember the "Shaky Tables" analogy. Errors always stack up! Practice a few problems involving \(X_1 + X_2\) versus \(2X\), and you'll be an expert in no time.