Using MGFs to Find Moments
You'll learn how to extract any moment of a random variable by taking derivatives of its moment generating function.
Now that you know what a moment generating function (MGF) is, it's time to see why it's so powerful: it encodes all the moments of a random variable in a single function, and we can extract them using calculus.
The Core Idea
Recall that the MGF is defined as M(t) = E[e^(tX)]. The name "moment generating" comes from this remarkable property:
The k-th derivative of M(t), evaluated at t = 0, gives you the k-th moment of X.
In symbols:
- M'(0) = E[X] (the first moment, or mean)
- M''(0) = E[X²] (the second moment)
- M^(k)(0) = E[X^k] (the k-th moment)
Why This Works
Think of the Taylor series expansion of e^(tX) around t = 0:
e^(tX) = 1 + tX + (t²X²)/2! + (t³X³)/3! + ...
When you take the expectation:
M(t) = 1 + tE[X] + (t²/2!)E[X²] + (t³/3!)E[X³] + ...
Each derivative "peels off" one layer: differentiating once gives E[X] + tE[X²] + ..., and evaluating at t = 0 isolates E[X].
Finding Variance
Once you have M'(0) and M''(0), you can find variance using the shortcut formula:
Var(X) = M''(0) - [M'(0)]²
This is just E[X²] - (E[X])².
Key Takeaway: Differentiate the MGF k times and evaluate at t = 0 to get the k-th moment—no integration needed, just calculus on a single function.