Expected Value of Election Results

AI Thread Summary
The discussion revolves around calculating the expected value of election outcomes for Party A winning the Senate and House. Initial probability calculations were incorrect due to the assumption of independence between the events. A Venn diagram helped clarify the relationships, leading to revised probabilities and an expected value of 1.1. The conversation also touched on issues with LaTeX rendering, particularly regarding overbars in the notation. The final consensus is that the expected value can be computed using linearity, regardless of correlations between the events.
vparam
Messages
17
Reaction score
3
Homework Statement
Suppose that in the 2050 election, polls show Party A has a 60% chance of winning the senate and a 50% chance of winning the house, and that if Party A wins the house, it has an 80% chance of winning the senate. Let X be the random variable whose value is the the number of congressional houses won by Party A. (a) Find the expected value of X.
Relevant Equations
##p(S|H) = \frac{p(S \cap H)}{p(H)}##
##E(X)=\sum_{s \in S}{p(s)X(s)}##
I submitted this solution, and it was marked incorrect. Could I get some feedback on where I went wrong?

Let S represent the event that Party A wins the senate and H represent the event that Party A wins the house.

There are 4 cases: winning the senate and house (##S \cap H##), winning just the senate (##S \cap \overline{H}##), winning just the house (##H \cap \overline{S}##), and winning neither (##\overline{S} \cap \overline{H}##).

I computed the probabilities for each compound event as follows:
$$p(S \cap H) = p(S|H)p(H)= (.8)(.5) = .4$$ $$p(S \cap \overline{H}) = p(S)p(\overline{H}) = (.6)(.5) = .3$$ $$p(H \cap \overline{S})=(p(S) - p(S \cap H))p(H) = (.6 - .4)(.5) = .1$$ $$p(\overline{S} \cap \overline{H}) = p(\overline{S})p(\overline{H}) = (.4)(.5) = .2$$
This means that the expected value is:
$$E(X) = 2(.4) + 1(.3) + 1(.1) + 0(.2) = 1.2$$
 
Physics news on Phys.org
vparam said:
There are 4 cases: winning the senate and house (##S \cap H##), winning just the senate (##S \cap \overline{H}##), winning just the house (##H \cap \overline{S}##), and winning neither (##\overline{S} \cap \overline{H}##).
Your overbars seem to show up when I quote your post, but I can't see them in the original.
vparam said:
I computed the probabilities for each compound event as follows:
$$p(S \cap H) = p(S|H)p(H)= (.8)(.5) = .4$$ $$p(S \cap \overline{H}) = p(S)p(\overline{H}) = (.6)(.5) = .3$$ $$p(H \cap \overline{S})=(p(S) - p(S \cap H))p(H) = (.6 - .4)(.5) = .1$$ $$p(\overline{S} \cap \overline{H}) = p(\overline{S})p(\overline{H}) = (.4)(.5) = .2$$
You have a total probability of winning the Senate as ##0.7##. And, you've assumed the probabilities are indepenent in some of those calculations. I drew a Venn diagram.
vparam said:
This means that the expected value is:
$$E(X) = 2(.4) + 1(.3) + 1(.1) + 0(.2) = 1.2$$
Two of those numbers are wrong.
 
  • Like
Likes vparam
vparam said:
$$p(S \cap \overline{H}) = p(S)p(\overline{H}) = (.6)(.5) = .3$$
Only if S and H are independent events. But it is easy to show that they are not independent.
vparam said:
$$p(\overline{S} \cap \overline{H}) = p(\overline{S})p(\overline{H}) = (.4)(.5) = .2$$
Again, only if S and H are independent events.
 
  • Like
Likes vparam
PeroK said:
I drew a Venn diagram.
Thank you for the reminder to use a Venn diagram! The problem breaks down more easily after drawing one out.

I've redone the calculations as follows:
$$p(S \cap H) = p(S|H)p(H)= (.8)(.5) = .4$$ $$p(S \cap \overline{H}) = p(S) - p(S \cap H) = .6 - .4 = .2$$ $$p(H \cap \overline{S}) = p(H) - p(S \cap H) = .5 - .4 = .1$$ $$p(\overline{S} \cap \overline{H}) = 1 - p(S \cup H) = 1 - (p(S) + p(H) - p(S \cap H)) = 1 - (.6 + .5 - .4) = .3$$
This means that the expected value is:
$$E(X) = 2(.4) + 1(.1) + 1(.2) + 0(.3) = 1.1$$
FactChecker said:
Only if S and H are independent events. But it is easy to show that they are not independent.

Again, only if S and H are independent events.
And thank you for pointing this out. This explains why my previous calculations don't make sense.

PeroK said:
Your overbars seem to show up when I quote your post, but I can't see them in the original.
Is \overline the standard code for overbars, or should I use something else?

Thank you both for the help!
 
vparam said:
Thank you for the reminder to use a Venn diagram! The problem breaks down more easily after drawing one out.

I've redone the calculations as follows:
$$p(S \cap H) = p(S|H)p(H)= (.8)(.5) = .4$$ $$p(S \cap \overline{H}) = p(S) - p(S \cap H) = .6 - .4 = .2$$ $$p(H \cap \overline{S}) = p(H) - p(S \cap H) = .5 - .4 = .1$$ $$p(\overline{S} \cap \overline{H}) = 1 - p(S \cup H) = 1 - (p(S) + p(H) - p(S \cap H)) = 1 - (.6 + .5 - .4) = .3$$
This means that the expected value is:
$$E(X) = 2(.4) + 1(.1) + 1(.2) + 0(.3) = 1.1$$

And thank you for pointing this out. This explains why my previous calculations don't make sense.Is \overline the standard code for overbars, or should I use something else?

Thank you both for the help!
I don't understand why the Latex renders only when I quote your post and not in the original. I can't see the overbars in your posts, but I can when I reply.

PS it doesn't work when I take them out of the quote box!

$$p(S \cap \overline{H}) = p(S) - p(S \cap H) = .6 - .4 = .2$$
 
  • Wow
Likes FactChecker
PeroK said:
I don't understand why the Latex renders only when I quote your post and not in the original. I can't see the overbars in your posts, but I can when I reply.

PS it doesn't work when I take them out of the quote box!

$$p(S \cap \overline{H}) = p(S) - p(S \cap H) = .6 - .4 = .2$$
I'm not sure why that's the case since I'm very new to Latex and it seems to show up properly on my end. I wrote up everything again using an alternate notation instead for visibility.

Probability calculations:
$$p(S \cap H) = p(S | H) p(H)= (.8)(.5) = .4$$ $$p(S \cap H^{\mathrm{C}}) = p(S) - p(S \cap H) = .6 - .4 = .2$$ $$p(H \cap S^{\mathrm{C}}) = p(H) - p(S \cap H) = .5 - .4 = .1$$ $$p(S^{\mathrm{C}} \cap H^{\mathrm{C}}) = 1 - p(S \cup H) = 1 - (p(S) + p(H) - p(S \cap H)) = 1 - (.6 + .5 - .4) = .3$$

Expected value computation:
$$E(X) = 2(.4) + 1(.1) + 1(.2) + 0(.3) = 1.1$$
 
1.1 is right. Good job working out the probabilities. Also, his is a trick question!

Expectancies add linearly. Let ##S## be the number of senate seats won, and ##H## the number of house seats won (each is either 0 or 1. This is called an indicator variable and is a very common idea to use)

Then you want to compute ##E(S+H)=E(S)+E(H)##. The correlations are irrelevant.
 
  • Like
Likes vparam and PeroK
PeroK said:
I don't understand why the Latex renders only when I quote your post and not in the original. I can't see the overbars in your posts,
Same here.
PeroK said:
but I can when I reply.

PS it doesn't work when I take them out of the quote box!

$$p(S \cap \overline{H}) = p(S) - p(S \cap H) = .6 - .4 = .2$$
When I reply to you, the overline of the first H shows.

$$p(S \cap \overline{H}) = p(S) - p(S \cap H) = .6 - .4 = .2$$
And now it doesn't. (There is an overline of the first H. If it matters, I am using the Firefox browser.)
 
Back
Top