In a population of $N$ families, $50\%$ of the families have three children, $30\%$ of the families have two children and the remaining families have one child. What is the probability that a randomly picked child belongs to a family with two children?

  1. $3/23$
  2. $6/23$
  3. $3/10$
  4. $3/5$

My attempt :

Using bayes theorem :

Required probability is

$= \frac{1/3*0.3}{1/3*0.5+1/3*0.3+1/3*0.2} = \frac{3}{10}$

But, somewhere answer is given $\frac{6}{23}$

Can you explain in formal way, please ?

  • 2
    $\begingroup$ Although the example excludes this, in reality some positive percentage of families has no children at all. What is the probability that a random child comes from such a family? Now modify your example to include 10% of childless families (remove them from the 50%, say) and do your Bayesian computation for the probability of a childless family. Compare. $\endgroup$ – Marc van Leeuwen Nov 18 '15 at 14:01
  • $\begingroup$ Where did the 1/3 come from in your formula? $\endgroup$ – Rick Nov 18 '15 at 17:49

Suppose there are ten families. Then five families have three children, which is 15 children; three families have two, which is six more; and the other two families have one child each, for a total of 23 children.

| cite | improve this answer | |
  • $\begingroup$ Yes , that's correct ,but what's wrong with my approach ? $\endgroup$ – ً ً Nov 18 '15 at 11:48
  • 2
    $\begingroup$ You have made the families equally likely, rather than the children equally likely. If the children are equally likely, then a three-child family is three times more likely to be chosen than a one-child family. $\endgroup$ – Empy2 Nov 18 '15 at 11:51
  • $\begingroup$ Yes. I got it thanks for explanation. $\endgroup$ – ً ً Nov 18 '15 at 12:21

Let $N$ be the number of families. Then $50\%$ of N have $3$ children, which means that there are $50\%N\cdot 3=\frac{15}{10}N$ children in these families and then $30\%N\cdot 2=\frac{6}{10}N$ children in families with $2$ children and $20\%N\cdot1=\frac{2}{10}N$ children in families of $1$ child. This makes a total of $$N(50\%\cdot3+30\%\cdot2+20\%\cdot1)=\frac{23}{10}N$$ children. So now pick a child. $$P(A_2)=\frac{\frac{6}{10}N}{\frac{23}{10}N}=\frac{6}{23}$$ where $A_2$ denotes the required event that the child belongs to a family with $2$ children.

| cite | improve this answer | |
  • $\begingroup$ How do you find $\frac 6 {10}$N? $\endgroup$ – Deliss Nov 18 '15 at 12:14
  • $\begingroup$ $30\%\cdot2N= \frac{30}{100}\cdot 2N=\frac{60}{100}N$ $\endgroup$ – Jimmy R. Nov 18 '15 at 12:16

Multiply the percents of families by N to get number of families:

  • 3 Children: 0.5*N families
  • 2 Children: 0.3*N families
  • 1 Child: 0.2*N families

Then, multiply the number of families by the number of children in each family:

  • 1.5*N children in three-child families
  • 0.6*N children in two-child families
  • 0.2*N children in one-child families

There are (1.5 + 0.6 + 0.2) * N = 2.3*N children total, of which 0.6*N come from two-child families. This reduces to 6/23.

| cite | improve this answer | |

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.