I have several statements related to the probability of three events and am trying to determine if the statements are true or false. For some of the answers, I have intuition but I would prefer to learn a provable answer.

  1. I have 3 events $E_1, E_2, E_3$ in a probability space. If $E_3 \subset E_1$ and $E_1, E_2$ are pairwise independent, are $E_2, E_3$ pairwise independent?

  2. $E_1, E_2, E_3$ are three jointly independent events. The, the events $E_1 \cup E_2$ and $E_3$ are pairwise independent.

  3. $E_1, E_2, E_3$ be three pairwise independent events in a probability space. Then the events $E_1\cup E_2$ and $E_3$ are pairwise independent.

My solutions thus far.

  1. My thought is it depends on the amount of space $E_3$ occupies of $E_1$ because independence is defined as $P(AB)=P(A)P(B)$.

  2. True. Jointly independent events are always pairwise independent.

  3. I am unsure about how to tackle this one.


For (1), pick $E_2,E_3$ your favorite dependent events and let $E_1$ be the sample space itself, the sure event. $E_1$ trivially contains $E_3$ and further $E_1$ is trivially independent with any other event.

For (2), this is a fair bit of algebra but fairly straightforward to prove. Just remember how intersection distributes over union, apply inclusion-exclusion and properties of independence, and factor.

$P((E_1\cup E_2)\cap E_3) = P((E_1\cap E_3)\cup (E_2\cap E_3))$ $ = P(E_1\cap E_3) + P(E_2\cap E_3) - P(E_1\cap E_2\cap E_3) $ $= P(E_1)P(E_3) + P(E_2)P(E_3) - P(E_1)P(E_2)P(E_3)$ $=P(E_3)\times (\cdots)$

For (3), consider an example where the events are pairwise but not mutually independent. For example considering the uniform distribution over the sample space $\{1,2,3,4\}$ and letting the event $E_i = \{i,4\}$. You have $E_1,E_2,E_3$ are each pairwise independent but not mutually independent.

Here we have $P(E_1\cup E_2)=\frac{3}{4}$, $P(E_3)=\frac{1}{2}$, and $P((E_1\cup E_2)\cap E_3) = \frac{1}{4}\neq \frac{3}{4}\times\frac{1}{2}$

  • $\begingroup$ Can you elaborate on 1. Also, how can i word my questions better so I dont get downvotes? $\endgroup$ – Avedis Dec 31 '18 at 18:44
  • 1
    $\begingroup$ @Avedis (1) is false by using as E_1 as the sure event and E_2,E_3 as any two dependent events. If you want an explicit example consider flipping a coin. E_1 is the event the coin is head or tail, E_2 is the event the coin is head, E_3 is the event the coin is tail. Note that E_1,E_2 are independent and E_3 is a subevent of E_1 though E_2,E_3 are clearly dependent on one another. $\endgroup$ – JMoravitz Dec 31 '18 at 18:51

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.