Detailed Explanation
This is a classic application of Bayes' theorem. We are given probabilities for the causes (the student either knows or guesses) and a probability for an observed effect (answering correctly). Bayes' theorem lets us reverse the direction of conditioning: instead of asking 'given that he knows, how likely is a correct answer?', we ask 'given that the answer is correct, how likely is it that he knew it?'. First we define events clearly, then use the multiplication theorem to find the probability of a correct answer through each path, and finally divide the 'knows' path by the total probability of a correct answer.
Formula Used
Bayes' theorem: \[P(E_1 \mid A) = \frac{P(E_1)\,P(A \mid E_1)}{P(E_1)\,P(A \mid E_1) + P(E_2)\,P(A \mid E_2)}\]
Solution
- Step 1: Define the events. Let \(E_1\) be the event that the student knows the answer and \(E_2\) be the event that the student guesses. Let \(A\) be the event that the student answers correctly. We want \(P(E_1 \mid A)\).
- Step 2: Write down the given probabilities. \[P(E_1) = \frac{3}{4}, \quad P(E_2) = \frac{1}{4}\] If the student knows the answer, it is certainly correct, so \[P(A \mid E_1) = 1\] If he guesses, he is correct with probability \[P(A \mid E_2) = \frac{1}{4}\]
- Step 3: Find the total probability of answering correctly using the multiplication theorem for each path and adding them. \[P(A) = P(E_1)P(A \mid E_1) + P(E_2)P(A \mid E_2)\] Substituting the values: \[P(A) = \frac{3}{4}\cdot 1 + \frac{1}{4}\cdot \frac{1}{4} = \frac{3}{4} + \frac{1}{16}\]
- Step 4: Simplify the total probability using a common denominator of 16. \[P(A) = \frac{12}{16} + \frac{1}{16} = \frac{13}{16}\]
- Step 5: Apply Bayes' theorem to reverse the conditioning. \[P(E_1 \mid A) = \frac{P(E_1)P(A \mid E_1)}{P(A)} = \frac{\frac{3}{4}\cdot 1}{\frac{13}{16}}\] This becomes: \[P(E_1 \mid A) = \frac{3}{4}\times \frac{16}{13} = \frac{48}{52} = \frac{12}{13}\]
Final Answer
The probability that the student knew the answer given that he answered correctly is \[P(E_1 \mid A) = \frac{12}{13}\]
Common Mistake
A common mistake is to stop at \(P(A \mid E_1) = 1\) and report the answer as \(1\), or to confuse \(P(E_1 \mid A)\) with \(P(A \mid E_1)\). Bayes' theorem reverses the conditioning, so you must divide the 'knows-and-correct' probability by the total probability of a correct answer \(\frac{13}{16}\), not just use the given \(P(A \mid E_1)\).