Two children, one boy

Ms. Jackson has two children and at least one of them is a boy. What's the probability both children are boys?

What do you think?
Ms. Jackson has two children and at least one is a boy. P(both are boys)?

The sample space

Write out every possibility for two children, where order matters (older child first):

Ω={(B,B),  (B,G),  (G,B),  (G,G)}\Omega = \{(B,B),\; (B,G),\; (G,B),\; (G,G)\}

Each outcome has probability 1/4. The condition "at least one boy" eliminates (G,G)(G,G), leaving three equally likely outcomes.

At least one boy

A mother with two children has at least one son. What's the probability both children are boys?

Sample space of two children:

BB
BG
GB
GG

GG (two girls) is eliminated. Of the 3 remaining, only BB has both boys → P = 1/3.

Conditioning on an event

When you condition on event BB, you restrict the sample space to outcomes where BB occurs and rescale probabilities so they sum to 1. If AA and BB are events, P(AB)=P(AB)/P(B)P(A|B) = P(A \cap B) / P(B).

Let AA = both children are boys, and BB = at least one child is a boy.

P(AB)=P(AB)P(B)=P({(B,B)})P({(B,B),(B,G),(G,B)})=1/43/4=13P(A|B) = \frac{P(A \cap B)}{P(B)} = \frac{P(\{(B,B)\})}{P(\{(B,B),(B,G),(G,B)\})} = \frac{1/4}{3/4} = \frac{1}{3}

A different question

Your colleague Ms. Parker also has two children. You see her walking with one of them, and that child is a boy. What's the probability both children are boys?

What do you think?
You see Ms. Parker with one child who is a boy. P(both children are boys)?

Why the answers differ

The two questions feel similar but they condition on different events.

The subtle difference
Part A: at least one boy\text{Part A: at least one boy}
This is about the family as a whole. You learn that the pair isn't GG, but you don't know which child is the boy. Three outcomes survive: BB, BG, GB.
Step 1 of 4
You see a boy

You see a colleague walking with one of her two children, and that child is a boy. What's the probability both children are boys?

Sample space of two children:

BB
BG
GB
GG

You saw one specific child who is a boy. The other child is independent: P(boy) = 1/2.

The difference is between "at least one of the two satisfies a property" (which creates correlation) and "this particular one satisfies a property" (which leaves the other independent). How you acquire information determines the conditional probability.

Practice

A family has 3 children, each equally likely to be a boy or girl. Given at least one is a boy, what's P(all three are boys)? (decimal like 0.14)
1/3