[Paper Review] On the so-called Boy or Girl Paradox

G. D’Agostini|arXiv (Cornell University)|Jan 5, 2010

Bayesian Modeling and Causal Inference3 references3 citations

TL;DR

This paper re-examines the Boy or Girl Paradox, demonstrating that the probability of a second child being a girl—given that one child is a girl with a specific name—does not depend on the name's rarity, provided identical names in the same family are not allowed. The key insight is that what matters is not the name itself, but the ability to uniquely identify a child, which transforms the probability from 1/3 (for 'at least one girl') to 1/2 (for 'a specific girl'), resolving the apparent paradox through Bayesian updating and identification logic.

ABSTRACT

A quite old problem has been recently revitalized by Leonard Mlodinow's book The Drunkard's Walk, where it is presented in a way that has definitely confused several people, that wonder why the prevalence of the name of one daughter among the population should change the probability that the other child is a girl too. I try here to discuss the problem from scratch, showing that the rarity of the name plays no role, unless the strange assumption of two identical names in the same family is taken into account. But also the name itself does not matter. What is really important is `identification', meant in an acceptation broader than usual, in the sense that a child is characterized by a set of attributes that make him/her uniquely identifiable (`that one') inside a family. The important point of how the information is acquired is also commented, suggesting an explanation of why several people tend to consider the informations "at least one boy" and "a well defined boy" (elder/youngest or of a given name) equivalent.

Motivation & Objective

• To clarify the persistent confusion in the Boy or Girl Paradox, especially as rekindled by Mlodinow’s book.
• To challenge the widespread assumption that the rarity of a child’s name affects conditional probability in two-child families.
• To demonstrate that the critical factor is not the name, but the ability to uniquely identify a child (e.g., 'the girl named Florida'), which changes the probability from 1/3 to 1/2.
• To argue that the distinction between 'at least one girl' and 'a specific girl' is crucial and often misunderstood in probabilistic reasoning.
• To emphasize that contextual information—such as how information is acquired and expressed—plays a vital role in probability assessment beyond bare facts.

Proposed method

• Uses Bayesian updating to compare the odds of two girls versus one girl, given that a specific child (by name or attribute) is known to be a girl.
• Introduces a background condition $I_2$ that prohibits identical names in the same family, eliminating the need to consider name duplication.
• Applies the Bayes factor to update prior odds using the likelihood of observing a named girl in families with two girls versus one girl.
• Constructs a contingency table (Table 4) under $I_2$ to compute joint probabilities of gender and name, showing that the probability of two girls given a named girl is 1/2.
• Analyzes the role of identification: a child identified uniquely (e.g., 'my daughter Claudia') is not equivalent to 'at least one girl', which changes the sample space.
• Argues that real-world information acquisition—through speech, emphasis, or context—often makes 'this one' distinct from 'at least one', leading to different probability assessments.

Experimental results

Research questions

• RQ1 Does the rarity of a child’s name affect the probability that the other child is of the same gender in a two-child family?
• RQ2 Why do people often conflate the probability of 'at least one girl' with 'a specific girl'?
• RQ3 What is the correct probability that both children are girls, given that one is a girl named Florida (or any name)?
• RQ4 How does the method of information acquisition—especially identification through unique attributes—affect conditional probability in such puzzles?
• RQ5 Why does the inclusion of name rarity in solutions to the paradox lead to misleading conclusions?

Key findings

• The probability that the other child is a girl, given that one child is a girl with a specific name, is 1/2, not 1/3, under the assumption that no two children in the same family share the same name.
• The rarity of the name has no effect on the final probability; the result holds regardless of how common or rare the name is.
• The key factor is not the name itself, but the ability to uniquely identify a child, which transforms the problem from 'at least one' to 'this specific one'.
• The Bayes factor approach confirms that the odds of two girls versus one girl are updated from 1:2 to 1:1 when a named girl is observed, resulting in equal posterior probabilities of 1/2.
• The confusion between 'at least one girl' and 'a specific girl' arises from real-world communication patterns, where names and emphasis serve as identification cues not captured in textbook formulations.
• The paper concludes that contextual and linguistic cues—such as emphasis or reference to a particular child—must be factored into probability assessment, as they fundamentally alter the sample space.