In: Math
Reminder: You obtain a positive test result for HIV. There is no
reason to believe that you should have a higher prior probability
of being HIV positive than the average the average person in
Australia. In Australia, about 30,000 people out of 24 million
people are HIV positive. The test has a false negative rate of 0.2%
(i.e., the probability of obtaining a negative result for a person
who is HIV positive is 0.002) and a false positive rate of 2.5%
(i.e., the probability of obtaining a positive result for a person
who is HIV negative is 0.025). After obtaining this test result,
what are the posterior odds in favour of you being HIV
positive?
A. 0.001 (this corresponds to odds of about 1 to 911 that you are
HIV positive)
B. 0.015 (this corresponds to odds of about 1 to 67 that you are
HIV positive)
C. 0.063 (this corresponds to odds of about 1 to 16 that you are
HIV positive)
D. 0.072 (this corresponds to odds of about 1 to 14 that you are
HIV positive)
E. 0.050 (this corresponds to odds of about 1 to 20 that you are
HIV positive)
Let HIV be the event that a random person is HIV positive. Let N and P be the event that the person tested negative and positive respectively.
Probability that a random person is HIV positive, P(HIV) = 30000 / 24 x 106 = 0.00125
Probability that a random person is not HIV positive (HIV negative), P(~HIV) = 1 - 0.00125 = 0.99875
Probability of obtaining a negative result for a person who is HIV positive = P(N | HIV) = 0.002
Probability of obtaining a positive result for a person who is HIV negative = P(P | ~HIV) = 0.025
Probability of obtaining a positive result for a person who is HIV positive = P(P | HIV) = 1 - P(N | HIV) = 1 - 0.002 = 0.998
Probability of obtaining a positive result = P(P)
= P(HIV) P(P | HIV) + P(~HIV) P(P | ~HIV) (By law of total probability)
= 0.00125 * 0.998 + 0.99875 * 0.025
= 0.02621625
Probability of being HIV positive given positive results = P(HIV | P)
= P(P | HIV) * P(HIV) / P(P) (By Bayes theorem)
= 0.998 * 0.00125 / 0.02621625
= 0.04758499
Posterior odds in favour of you being HIV positive = 0.04758499 / (1 - 0.04758499) = 0.05
E. 0.050 (this corresponds to odds of about 1 to 20 that you are HIV positive)