If a random variable X assumes the value of a Bernoulli(1/3) random variable with a probability of 1/4, and assumes the value of a Uniform(0,1) random variable with a probability of 3/4, how does one obtain a moment generating function of X?

I'm fuzzy on how exactly mixed distributions work, and I'm not sure that term even applies here, however I came across this problem and thought it to be like no other.

In any case, I believe the moment generating function for the uniform part of this variable would be ((e^t)-1)/t. And the Bernoulli mgf would be 2/3+(1/3)(e^t).

Given that the Bernoulli variable occurs with probability 3/4 and the Uniform variable with probability 1/4, how does one combine these into one cohesive mgf? Have I simply forgotten a formula?


The moment generating function is an expectation value. Particularly, it's $E(e^{tX}),$ so the question is how to take this expectation value when $X$ is a mixture of two distributions.

But you can think of a mixture as an initial coinflip which determines which conditional distribution the variable comes from. So by total probability we have $$E(e^{tX}) = E(e^{tX}|Bernoulli)P(Bernoulli) + E(e^{tX}|Uniform)P(Uniform).$$

In other words, it's the weighted sum of the two MGFs.

| cite | improve this answer | |
  • $\begingroup$ Thank you, I suspected it might be something this simple, but wasn't sure. $\endgroup$ – Bad at algebra and proofs Jun 16 '17 at 23:10
  • $\begingroup$ Yeah, but you're right to not assume... a lot of times it isn't. $\endgroup$ – spaceisdarkgreen Jun 16 '17 at 23:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.