Sum of Independent Normal Variables
You'll learn: When you add two (or more) independent normal random variables together, the result is also a normal random variable—and you can predict its mean and variance using simple rules.
The Core Idea
Imagine you're measuring two independent processes, both normally distributed: daily temperature fluctuation (mean 0°C, std dev 2°C) and measurement error from your thermometer (mean 0°C, std dev 1°C). When you combine these effects, what distribution do you get?
Great news: the sum is also normal. Even better, the parameters follow intuitive rules:
- Mean of the sum = Sum of the means
- Variance of the sum = Sum of the variances (for independent variables)
Why This Matters
This property makes normal distributions incredibly powerful in data science. When multiple independent normal factors combine (measurement errors, random fluctuations, aggregated data), you don't need complex calculations—the result stays in the normal family.
Example
Suppose:
- X ~ Normal(μ₁ = 10, σ₁² = 4)
- Y ~ Normal(μ₂ = 5, σ₂² = 9)
- X and Y are independent
Then:
- Z = X + Y is also Normal
- μ_Z = 10 + 5 = 15
- σ²_Z = 4 + 9 = 13 (so σ_Z = √13 ≈ 3.6)
Notice you add variances, not standard deviations. This reflects how uncertainty compounds when combining independent sources.
The Magic Word: Independent
This only works cleanly when variables are independent. If they're correlated, the variance formula becomes more complex (we'll save that for later!).
Key Takeaway: The sum of independent normal variables is normal, with mean equal to the sum of means and variance equal to the sum of variances—making combined normal processes easy to analyze.