Question
A family has two children whose genders and birth order are unknown.
(1) Given that the older child is a girl, find the probability that both children are girls.
(2) Given that one of the children is a girl, find the probability that both children are girls.
(3) Upon observing one child who turns out to be a girl, find the probability that both children are girls.
(4) Given that at least one child is a girl, find the probability of observing a girl when one child is selected at random.
Step-by-step solution
Let B denote a boy and G denote a girl. For a family with two children, disregarding gender and birth order, the sample space is Ω = {(B, B), (B, G), (G, B), (G, G)}, where the first letter represents the older child and the second letter represents the younger child. Each outcome is equally likely with probability . (1) Let event A be "the older child is a girl." Then the sample points in A are {(G, B), (G, G)}. Let event B be "both children are girls." Then the sample points in B are {(G, G)}. The desired probability is the conditional probability .
(2) Let event C be "one of the children is a girl," meaning at least one child is a girl. Then the sample points in C are {(B, G), (G, B), (G, G)}. Let event B be "both children are girls." Then the sample points in B are {(G, G)}. The desired probability is the conditional probability .
(3) "Observing one child who turns out to be a girl" is an observation event; denote it as event S. Let the events corresponding to family compositions GG, GB, BG, BB be , respectively. Under each family composition, the probability of observing a girl is as follows: By the law of total probability, the probability of observing a girl is: The desired probability is . By Bayes' theorem:
(4) This problem asks for the probability of "observing a girl" (event S) given that "at least one child is a girl" (event C), i.e., . By the definition of conditional probability: If event S occurs (a girl is observed), then event C (at least one child is a girl) necessarily occurs. Therefore . From the computations in (2) and (3), and .
Final answer
(1) (2) (3) (4)
Marking scheme
The following rubric is based on the official solution.
1. Checkpoints (max 7 pts total)
- Part (1) (1 pt)
- Correct result: . [1 pt]
- *(Note: If only the answer is given without working, full credit is awarded for this sub-part; if the reasoning is clearly flawed but the answer happens to be correct, no credit is given.)*
- Part (2) (2 pts)
- Key identification: Correctly identifies that the conditioning event "at least one girl" corresponds to a reduced sample space of size 3 (i.e., ), or explicitly states . [1 pt]
- Correct result: . [1 pt]
- Part (3) (2 pts)
- Observation model / Law of total probability: The computation reflects the reasoning that in a mixed-gender family (GB/BG) the probability of observing a girl is , or the law of total probability is correctly applied to obtain . [1 pt]
- *(Note: If the student uses an independence argument, i.e., argues that "observing one child to be a girl does not affect the gender distribution of the other child," and the logic is sound, this point is awarded.)*
- Correct result: . [1 pt]
- Part (4) (2 pts)
- Logical derivation: Identifies that (i.e., "observing a girl" necessarily implies "at least one child is a girl") and thereby concludes ; or correctly constructs the ratio . [1 pt]
- Correct result: . [1 pt]
- Total (max 7)
2. Zero-credit items
- Merely copying the conditional probability definition without any concrete numerical substitution or set analysis earns 0 pts.
- An answer of for Part (2) (confusing ordered vs. unordered outcomes, or conflating Parts (1) and (2)) earns 0 pts for that part.
- An answer of for Part (3) (ignoring the observation probability and incorrectly assuming that having a girl guarantees she is observed) earns 0 pts for that part.
- An answer of for Part (4) (incorrectly treating "at least one child is a girl" and "observing a girl" as the same event) earns 0 pts for that part.
3. Deductions
- Arithmetic errors: Pure computational mistakes (e.g., incorrect fraction division) incur a -1 pt deduction.
- Carry-forward errors:
- If the logical formula in Part (4) is correct (e.g., ) but an incorrect numerical value from Part (2) or (3) is carried forward, no additional deduction is applied (only the points lost in the earlier parts are forfeited); credit is given for correct logic.
- Missing reasoning:
- In Parts (3) or (4), if events are not defined or probability sources are not stated, and the work consists of unexplained numerical expressions rendering the logic unreadable, a -1 pt deduction applies.
- If an obvious probability fallacy appears, such as or , a -1 pt deduction applies.