0
$\begingroup$

I have a statistics problem and, although I have solved most of it, I am stuck for quite some time in the last question.

Say that there we have a population where the analogy of men to women is θ. Suppose that the probability of a woman being colorblind is $q^2$ and the probability of a man being colorblinf $q$.

Up to now I have found, if choosing one random person, the probability of him/her being colorblind and also that probability of a person being a man, given that he is colorblind. However, now I can't find a way to approach this question:

Suppose we choose two people out of the population radomly. What is the probability that at least one person is colorblind?

I was thinking of approaching it by finding the probability of none being colorblind and then finding its compliment, but I can't figure out how to do this, given that we have a different percentage for men and women. Any help on this would be appreciated. Thank you!

$\endgroup$
  • $\begingroup$ To be clear, there are $\theta$ men for every $1$ woman, correct? $\endgroup$ – Noble Mushtak Dec 28 '18 at 22:20
  • 1
    $\begingroup$ No, what I mean is P(A)/ P(A') = θ , where A is that the subject is male. (So I guess, men/women= θ) $\endgroup$ – butterflyflyaway Dec 28 '18 at 22:22
  • $\begingroup$ Condition the complement of your event on the number of females sampled: 0, 1, or 2. $\endgroup$ – Laars Helenius Dec 28 '18 at 22:26
  • $\begingroup$ Why only on females though? $\endgroup$ – butterflyflyaway Dec 28 '18 at 22:27
2
$\begingroup$

Let there be $m$ men and $w$ women (unfortunately, I also have to assume there are also $0$ non-binary people in this population). We have the following:

$$\frac m w=\theta\rightarrow m=\theta w$$

Thus, the probabilities of being a man or a woman are:

$$P(man)=\frac{m}{m+w}=\frac{\theta w}{\theta w+w}=\frac{\theta}{\theta+1}$$ $$P(woman)=\frac{w}{m+w}=\frac{w}{\theta w+w}=\frac{1}{\theta+1}$$

Now, we are also given the following:

$$P(colorblind | man)=q^2$$ $$P(colorblind | woman)=q$$

Then, we can use the formula $P(A \cap B)=P(B)P(A|B)$ to find the following:

$$P(colorblind \cap man)=P(man)P(colorblind | man)=\frac{q^2\theta}{\theta+1}$$ $$P(colorblind \cap woman)=P(woman)P(colorblind | woman)=\frac{q}{\theta+1}$$

Finally, since man and woman are the only two cases in this population, $woman=man^C$—that is, woman is the complement case of man. Thus, we can use the formula $P(A)=P(A \cap B)+P(A \cap B^C)$ to find:

$$P(colorblind)=P(colorblind \cap man)+P(colorblind \cap woman)=\frac{q^2\theta+q}{\theta+1}$$

And of course, to find the probability is not colorblind, we use $P(A^C)=1-P(A)$:

$$P(colorblind^C)=1-P(colorblind)=\frac{(1-q^2)\theta+1-q}{\theta+1}$$

Now, the probability that at least one person is colorblind has three cases in it:

  • Both people are colorblind -> $P(colorblind)^2$
  • First person is colorblind, second person is not -> $P(colorblind)\cdot P(colorblind^C)$
  • Second person is colorblind, first person is not -> $P(colorblind)\cdot P(colorblind^C)$

Thus, to find the total probability of there being one color blind person, we just add up all of these cases:

$$P(at\ least\ one\ colorblind)=P(colorblind)^2+2P(colorblind)\cdot P(colorblind^C) \\=\frac{(q^2\theta+q)^2+2(q^2\theta+q)(\theta(1-q^2)+1-q)}{(\theta+1)^2}$$

I'll leave doing the algebra to you. Good luck!

$\endgroup$
  • 1
    $\begingroup$ To complete the problem, assume the sampling is independent. Then $P(\text{at least one colorblind})=1-P(\text{not colorblind})^2=1-[1-P(\text{colorblind})]^2$ $\endgroup$ – M. Nestor Dec 28 '18 at 22:31
  • $\begingroup$ One more question. This was the answer that I found when I was picking one person and finding the probability of whether he/she is colorblind. So, it remains the same even after I choose two people? $\endgroup$ – butterflyflyaway Dec 28 '18 at 22:33
  • $\begingroup$ @butterflyflyaway Sorry, see my new edited answer for the actual, final answer of the probability of at least one person being color blind. $\endgroup$ – Noble Mushtak Dec 28 '18 at 22:36
  • $\begingroup$ @M.Nestor This is a valid solution, but it is equivalent to $1-(1-p)^2=1-(p^2+1-2p)=2p-p^2=p^2+2p-2p^2=p^2+2p(1-p)$, which is what I did. $\endgroup$ – Noble Mushtak Dec 28 '18 at 22:38

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.