Interview Question

Finance Operations Analyst Interview

-Mountain View, CA

Google

Pretend 1% of the population has a disease. You have a test that determines if you have that disease, but it's only 80% accurate and 20% of the time you get a false positive, how likely is it you have the disease.

AnswerAdd Tags

Interview Answers

8 Answers

3

1%. The first sentence gives you all the info you need.

Bryan on

1

1% chance if using global population. Chances would decrease or increase depending on certain variable or demographic factors so I would say you need more data to give a specific answer. Things like age , gender , race , need to be taken into consideration if you are a 33 year old male and the subject disease is ovarial cancer then you have 0 percent chance . If you are a 40 year old nun and the disease is The clap you have less probably of being infected than say a 26 year old prostitute ..

Anonymous on

2

Sample: 1,000 X=10 P=.8 If the test is only 80% accurate then 8 of the 10 infected will be positive and 2 will walk around unbeknownst, but 20% of the non-infected population will be falsely identified, 1000-10=990 (non-infected) 990*.2=198 (false positive) + 8 (true positive)= 296 (total positive tests) Of this population only 8/296=2.7% were correctly identified. This is the chance you have the disease

Anonymous on

12

You don't need the test. The answer is given... 1% of the population has the disease... Answer: 1%

Thanasis on

6

1% Given: 1. 1% of the population has a disease 2. A test exists to identify disease carriers The questions makes no statement regarding whether or not you have taken the test, the second piece of data has no relevance. You have a 1% chance

Read the question on

1

Its easier if you draw this out in quadrants, Accuracy = 80% = P(Test shows true and having disease) + P (Test showing false and not having disease) Also, we know false positive rate = P(Test shows true but not having disease) = 20% That leaves us with the fact that P(Test being false and person having disease) = 0% Hence likelihood of having the disease is 1%

Anonymous on

1

divide by reciprocal, multiply by .8% = 1%

Anonymous on

1

I thinks this question is ambiguous. If it meant the possibility that you have the disease, then 1%; If it meant the possibility that your test result shows you have the disease, then 1%*80%+99%*20%

Anonymous on

Add Answers or Comments

To comment on this, Sign In or Sign Up.