Multinomial Distribution Explained with Worked Examples

In independent trials with outcomes of probabilities , the probability of exactly of each is .

Write 2x or 2*x to multiply, ^ for powers, and ; between simultaneous equations. You can also type factor, expand, derivative of, integrate … from a to b, limit … as x->0 or 15% of 80.

Try an example

Before you start

This page is easier if you are comfortable with:

What Multinomial distribution means

Rolling a fair die 6 times, the chance each face appears once is . For 10 draws with , the probability of counts is . With 4 trials and four equally likely outcomes, the counts can occur in arrangements, each with probability , giving .

The multinomial coefficient counts the arrangements (permutations of repeated objects, stars and bars); with it is the binomial distribution (the binomial distribution). Each count is binomial with mean and variance , and the counts are negatively correlated, (covariance).

The distribution underlies the chi-squared test for categorical data (the chi-squared test statistic), the likelihood of word counts in text models and the genetics of multiple alleles. Its likelihood is maximised by relative frequencies (maximum likelihood estimation).

What is it used for?

The multinomial distribution extends the binomial to experiments with more than two outcomes.

How to use multinomial distribution

  1. Check that the counts add to and the probabilities to 1.
  2. Compute the multinomial coefficient .
  3. Multiply by .
  4. Evaluate.

Graph

Probabilities of the three multinomial exampleseach face once: 0.0154(5,3,2): 0.0851(2,1,1,0): 0.046900.511.522.533.5400.020.040.060.080.1Probabilities of the three multinomial examplesexampleyGraph with x from 0 to 4 and y from 0 to 0.1. Shows bars (3); point each face once: 0.0154 at (1, 0.01543); point (5,3,2): 0.0851 at (2, 0.08505); point (2,1,1,0): 0.0469 at (3, 0.0469).

The multinomial coefficient counts the ways each pattern can occur.

Worked examples

These examples are worked by hand, step by step, so you can compare your own working.

Example 1: each face once

Find the probability that 6 rolls of a fair die show every face once.

Worked example
Answer
0.015430.01543
  1. Formula

Example 2: counts (5, 3, 2)

Find the probability of counts in 10 draws with .

Worked example
Answer
0.085050.08505
  1. Coefficient
  2. Probability

Example 3: counts (2, 1, 1, 0)

Find the probability of counts in 4 trials with four equally likely outcomes.

Worked example
Answer
0.04690.0469
  1. Coefficient
  2. Probability

Special cases

A few situations come up again and again. Knowing them saves time and helps you spot answers that cannot be right.

n = 6, p = (0.5, 0.3, 0.2), counts (3, 2, 1)

coefficient 6!/(3!2!1!)

60.

Probability term

0.5³ × 0.3² × 0.2

0.00225.

Total

60 × 0.00225

0.135.

Mean of category 1

np₁

3.

Word problems

In n independent trials with k outcomes, P(n₁, …, n_k) = n!/(n₁!…n_k!) p₁^(n₁)…p_k^(n_k). Each count is binomial with mean npᵢ.

Coefficient

Find 6!/(3!2!1!).

  1. 720/12
  2. Check: ✓60

Probability term

Find 0.5³ × 0.3² × 0.2.

  1. 0.125 × 0.09 × 0.2
  2. Check: ✓0.00225

Total probability Higher

Multiply the coefficient and the term.

  1. 60 × 0.00225
  2. Check: ✓0.135

Mean Higher

Find the expected count of outcome 1.

  1. 6 × 0.5
  2. Check: ✓3

Common mistakes

Forgetting the coefficient

only the product ✗

multiply by n!/(∏nᵢ!) ✓

Orderings.

Counts not summing to n

3 + 2 + 2 ✗

sum to n ✓

Check.

Probabilities not summing to 1

check ✗

Σpᵢ = 1 ✓

Valid distribution.

Independent categories

independent counts ✗

negatively correlated ✓

Fixed total.

Exam tips

  1. n!/(∏nᵢ!).
  2. ∏pᵢ^nᵢ.
  3. Σnᵢ = n.
  4. Mean npᵢ.
  5. Binomial if k = 2.

Practice questions

Try each one on paper, then open the answer. Use the solver above to see the full working.

1. 4!/(2!1!1!) Foundation

12

2. 5!/(2!2!1!) Foundation

30

3. n = 3, p = 1/3 each, counts (1,1,1) Foundation

0.2222

4. Mean n = 20, p = 0.25 Higher

5

5. n = 4, p = (0.5, 0.5), counts (2, 2) Higher

0.375

6. 6!/(2!2!2!) Higher

90

Practice generator

Unlimited fresh questions with instant marking. Choose a level, work it out on paper, then type your answer.

Solve: …

Key terms

Multinomial distribution
Counts across k categories.
Multinomial coefficient
n!/(n₁!…n_k!).
Category probability (pᵢ)
Chance of outcome i.
Trial
One independent experiment.
Binomial
Two categories.
Factorial
n!.

Frequently asked questions

What is the multinomial distribution?

The generalisation of the binomial to k outcomes.

What is the formula?

n!/(∏nᵢ!) ∏pᵢ^nᵢ.

What is the mean of each count?

npᵢ.

What if k = 2?

It is the binomial.

Which page covers the hypergeometric?

hypergeometric distribution.

Which page covers the Poisson process?

Poisson process.

More in Advanced statistics and stochastic processes and the formula sheets.