$20\%$ of a class earned an $A$ in the mid term exam. Of those $20\%$, $50\%$ of them also scored an $A$ in the final exam. Of the people who scored less than $A$ in the mid-term, $20\%$ of them scored an $A$ in the final. If I pick up a random final paper with a score of $A$, what is the chance that the student also scored an $A$ in the mid term.

Ok so my calculations are: $10\%$ of students scored $A$ in mid term and final. $16\%$ of students scored A in final only. So, the chance that a student with a final $A$ score, who also got $A$ in mid term, in both is simply $$\frac{10}{10+16}=\frac{10}{26}$$ approximately $38\%$.

I'd like to think I am right but it seems a bit too easy which leads me to think I am overlooking something.

  • $\begingroup$ There is something strange.. you said $20\%$ at the beginning, about the ones who scored A in the mid term but you made the calculation with a $10\%$. What is the correct percentage? $\endgroup$ – Von Neumann Mar 31 '16 at 10:05
  • $\begingroup$ typo, sorrry! been edited $\endgroup$ – stackdsewew Mar 31 '16 at 10:13

You are correct. Formally, you can define the events:

  1. $A_m:=$ a student scored $A$ in the mid-term exam.
  2. $A_f:=$ a student scored $A$ in the final exam.

with given probabilities $P(A_m)=0.2, P(A_f\mid A_m)=0.5, P(A_f \mid A_m^c)=0.2$ and you are looking for $P(A_m\mid A_f)$. By Bayes Rule (and the law of total probability for the denominator) you get that \begin{align}P(A_m \mid A_f)&=\frac{P(A_f\mid A_m)P(A_m)}{P(A_f)}=\frac{P(A_f\mid A_m)P(A_m)}{P(A_f\mid A_m)P(A_m)+P(A_f\mid A_m^c)P(A_m^c)}\\[0.2cm]&=\frac{0.5(0.2)}{0.5(0.2)+0.2(0.8)}=\frac{10}{10+16}=\frac{10}{26}\end{align}

The only difference in your approach is that you calculated $P(A_f\cap A_m)$ from the formula $$P(A_f\cap A_m)=P(A_f\mid A_m)P(A_m)=0.2(0.5)=0.10$$ or $10\%$ and similarly $P(A_f^c \cap A_m)$. These calculations gave you $P(A_f)$ from the formula $$P(A_f)=P(A_f\cap A_m)+P(A_f\cap A_m^c)$$ which is of course perfectly ok!


Let's mind in this way: we define $T$ as the total amount of people who took the test. Hence $$20\% T = \frac{20}{100} T = A_{\text{mid term}} = \frac{1}{5} T$$ $$50\% (20\% T) = \frac{500}{10000} T = A_{\text{final}} = \frac{1}{10} T$$ $$20\% (80\% T) = \frac{1800}{10000} T = A_{\text{final}}^{\text{less than A in midterm}} = \frac{4}{25} T$$

Now the total amount of people who got $A$ in the final exam is

$$\frac{1}{10} T + \frac{4}{25} T = \frac{13}{50} T$$

To get the total amount of possible people who scored $A$ in both the test we just need to calculate

$$P = \frac{\frac{1}{10}}{\frac{13}{50}} = \frac{1}{10}\cdot \frac{50}{13} = \frac{5}{13}$$

Which is exactly your answer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.