Course contentsShow
← Mathematics
Lesson 2932 of 3,12158. Joint and Limit BehaviorFree lesson

Properties of Conditional Distributions

Explore how conditional distributions behave, including normalization and computing conditional expectations and variances.

Properties of Conditional Distributions

You'll learn: How conditional distributions behave mathematically, including their normalization property and how to compute conditional expectations and variances.

The Core Idea

When you condition on an event, you're essentially creating a new probability world restricted to that event. Just like ordinary distributions, conditional distributions must follow certain rules and have their own means and variances—but computed within this restricted world.

Think of it like examining only the players on a basketball team who are over 6 feet tall. The "conditional distribution" of weights among these tall players forms a valid distribution on its own, with its own average and spread.

Key Properties

1. Normalization

Every conditional distribution integrates (or sums) to 1 over its domain. For discrete variables: ∑xfX∣Y(x∣y)=1\sum_x f_{X|Y}(x|y) = 1

For continuous: ∫−∞∞fX∣Y(x∣y) dx=1\int_{-\infty}^{\infty} f_{X|Y}(x|y)\,dx = 1

This must hold for each fixed value of the conditioning variable.

2. Conditional Expectation

The conditional mean is the expected value computed using the conditional distribution: E[X∣Y=y]=∑xx⋅fX∣Y(x∣y)(discrete)E[X|Y=y] = \sum_x x \cdot f_{X|Y}(x|y) \quad \text{(discrete)} E[X∣Y=y]=∫−∞∞x⋅fX∣Y(x∣y) dx(continuous)E[X|Y=y] = \int_{-\infty}^{\infty} x \cdot f_{X|Y}(x|y)\,dx \quad \text{(continuous)}

3. Conditional Variance

The conditional variance measures spread within the conditioned world: Var(X∣Y=y)=E[X2∣Y=y]−(E[X∣Y=y])2\text{Var}(X|Y=y) = E[X^2|Y=y] - (E[X|Y=y])^2

Like regular variance, it uses the same computational shortcut but with conditional expectations.

Key Takeaway: Conditional distributions are valid probability distributions themselves, satisfying normalization and possessing their own expectations and variances computed exclusively within the conditioned scenario.