Law of Total Probability: Why It Works and How to Use It

Law of Total Probability

Submit Homework

The law of total probability is a foundational tool in mathematical statistics tutoring. It allows you to compute the probability of an event by breaking it into mutually exclusive cases. Students often use the formula mechanically without understanding why it works. This page explains the intuition, structure, and exam‑relevant applications.

If \(B_1, \dots, B_k\) form a partition, then \[ P(A) = \sum_{i=1}^k P(A \mid B_i) P(B_i). \]

Why does this work? Because every outcome in the sample space must fall into exactly one of the partition events \(B_i\). So the event \(A\) can be decomposed into disjoint pieces: \(A = (A \cap B_1) \cup \cdots \cup (A \cap B_k)\). Probabilities of disjoint events add, and each piece can be written using conditional probability. This decomposition is the backbone of Bayesian inference and many applied probability models.

1. Identify a partition of the sample space. Choose events \(B_1, \dots, B_k\) that are mutually exclusive and exhaustive.
2. Break the target event into pieces. Write \(A\) as the union of \(A \cap B_i\).
3. Use additivity. \(P(A) = \sum P(A \cap B_i)\).
4. Apply conditional probability. \(P(A \cap B_i) = P(A \mid B_i) P(B_i)\).
5. Combine the pieces. \(P(A) = \sum P(A \mid B_i) P(B_i)\).
6. Interpret the result. You are averaging conditional probabilities weighted by how likely each case is.

Suppose a medical test is used in two populations:

• High‑risk group: 20% of patients
• Low‑risk group: 80% of patients

Test positivity rates:

• High‑risk: \(P(\text{Positive} \mid \text{High}) = 0.30\)
• Low‑risk: \(P(\text{Positive} \mid \text{Low}) = 0.05\)

Then the overall positivity rate is:

\[ P(\text{Positive}) = 0.30(0.20) + 0.05(0.80) = 0.06 + 0.04 = 0.10. \]

So 10% of all patients test positive.

• Using events that do not form a partition.
• Forgetting that the \(B_i\) must be mutually exclusive.
• Mixing up conditional and unconditional probabilities.
• Applying the formula without checking independence (not required).

The law of total probability is essential for Bayesian inference, diagnostic testing, mixture models, and hierarchical models. It appears in nearly every graduate statistics exam and underlies the derivation of Bayes’ rule.

This idea connects directly to:

• Math Statistics & Probability Theory (parent spoke)
• Tutoring Services
• Question Hub
• Statistics Post Hub