AP Statistics Notes: Tree Diagrams and Conditional Probabilities

Definition: A tree diagram is a visual tool used to systematically think through conditional probabilities. It represents sequences of events as paths that resemble the branches of a tree.
Purpose: These diagrams are particularly useful for visualizing the sample space and calculating probabilities of compound events, especially when the probability of a later event depends on the outcome of an earlier one.
General Rule for Tree Diagrams: * Branches represent different possible outcomes of a stage. * The sum of probabilities for branches originating from the same point must equal $1.0$ . * Probabilities along a single path are multiplied to find the joint probability of that specific sequence of events. * To find the total probability of an event that occurs in multiple paths (disjoint events), the individual joint probabilities are added together (Addition Rule).

Study References: * Harvard School of Public Health: H. Wechsler, G. W. Dowdall, A. Davenport, and W. DeJong, "Binge Drinking on Campus: Results of a National Study." * American Journal of Health Behavior: Study regarding alcohol-related automobile accidents among drinkers aged $21$ to $34$ .
Definitions of Binge Drinking: * Men: Defined as consuming five or more drinks in a row. * Women: Defined as consuming four or more drinks in a row. * Reason for Difference: The threshold varies due to the average difference in body weight between men and women.
Population Statistics (College Students): * Engage in binge drinking ( $B$ ): $44\%$ * Drink moderately ( $M$ ): $37\%$ * Abstain entirely ( $A$ ): $19\%$
Accident Statistics (Among those aged $21$ to $34$ ): * Among binge drinkers ( $B$ ): $17\%$ have been involved in an alcohol-related automobile accident. * Among non-binge drinkers ( $NB$ ): Includes moderate drinkers and those who abstain. * The transcript specifies: $P(\text{accident} | \text{moderate}) = 0.09$ * The transcript specifies: $P(\text{accident} | \text{abstain}) = 0$

Stage 1: Drinking Category (Totaling $1.0$ or $100\%$ ): * $P(B) = 0.44$ * $P(M) = 0.37$ * $P(A) = 0.19$
Stage 2: Conditional Accident Probabilities: * From Binge Drinkers ( $B$ ): * $P(\text{accident} | B) = 0.17$ * $P(\text{no accident} | B) = 0.83$ * From Moderate Drinkers ( $M$ ): * $P(\text{accident} | M) = 0.09$ * $P(\text{no accident} | M) = 0.91$ * From Abstainers ( $A$ ): * $P(\text{accident} | A) = 0$ * $P(\text{no accident} | A) = 1.0$

Binge Drinking and Accident: * $P(B \cap \text{acc}) = (0.44) \times (0.17) = 0.075$
Binge Drinking and No Accident: * $P(B \cap \text{no acc}) = (0.44) \times (0.83) = 0.365$
Moderate Drinking and Accident: * $P(M \cap \text{acc}) = (0.37) \times (0.09) = 0.033$
Moderate Drinking and No Accident: * $P(M \cap \text{no acc}) = (0.37) \times (0.91) = 0.337$
Abstaining and Accident: * $P(A \cap \text{acc}) = (0.19) \times (0) = 0$
Abstaining and No Accident: * $P(A \cap \text{no acc}) = (0.19) \times (1) = 0.190$

What's the probability that a randomly selected student has had an alcohol-related car accident? * This is the sum of all joint probabilities ending in "accident." * $P(\text{acc}) = P(B \cap \text{acc}) + P(M \cap \text{acc}) + P(A \cap \text{acc})$ * $P(\text{acc}) = 0.075 + 0.033 + 0 = 0.108$
What's the probability that a randomly selected college student is a binge drinker and has had an alcohol-related car accident? * This is the joint probability established by following the first branch ( $B$ ) and then the accident branch. * $P(B \cap \text{acc}) = 0.075$
What's the probability that a student who has had an alcohol-related car accident is a binge drinker? * This is a conditional probability: $P(B | \text{acc})$ . * Calculation: $\frac{P(B \cap \text{acc})}{P(\text{acc})} = \frac{0.075}{0.108} \approx 0.694$
If a student has never had an alcohol-related accident, what is the probability they are actually a binge drinker? * This is a conditional probability: $P(B | \text{no acc})$ . * First, calculate $P(\text{no acc}) = 0.365 + 0.337 + 0.190 = 0.892$ . * Calculation: $\frac{P(B \cap \text{no acc})}{P(\text{no acc})} = \frac{0.365}{0.892} \approx 0.409$

Formula: Bayes's Rule provides a way to calculate a conditional probability by reversing the conditioning. * $P(B | A) = \frac{P(A | B)P(B)}{P(A | B)P(B) + P(A | B^C)P(B^C)}$
Note on Complexity: The transcript notes that while this formula can be used (offering "TB testing probabilities" as a potential challenge for "masochists"), it is often far easier and more intuitive to simply draw the tree diagram to reach the same result.