
Benjamin V. answered 07/26/22
Master's in Applied Statistics -- Collegiate Tutoring Experience
Applying Bayes' Theorem,
P(Seatbelt | Ejected) = (P(Ejected | Seatbelt)*P(Seatbelt)) / P(Ejected)
The denominator can be split conditionally, using the Law of Total Probability, based on whether or not the passenger was wearing a seatbelt, yielding
= (P(Ejected | Seatbelt)*P(Seatbelt)) / (P(Ejected | Seatbelt)P(Seatbelt) + P(Ejected | No Seatbelt)P(No Seatbelt))
A few things to notice: the numerator and the first term in the sum in the denominator are identical (P(Ejected | Seatbelt)P(Seatbelt)), as is common when applying Bayes' Theorem in this way. Secondly, I conditioned the way I did because it incorporates all of the known information.
The answer for part A is then
(.16 * .66) / (.16*.66 + .4 * (1 - .66)) = (.16 * .66) / (.16*.66 + .4 * .34)) = .4371
For part B, follow a similar approach:
P(No Seatbelt | Not Ejected) = (P(Not Ejected | No Seatbelt)*P(No Seatbelt)) / P(Not ejected)
= (P(Not Ejected | No Seatbelt)*P(No Seatbelt)) / (P(Not Ejected | No Seatbelt)*P(No Seatbelt) + P(Not Ejected | Seatbelt)*P(Seatbelt))
= ((1 - .4)*(1 - .66)) / ((1 - .4)*(1 - .66) + (1 - .16)*(.66))
= (.6 * .34) / (.6 * .34 + .84 * .66) = .2690