[blackermann]

I'm working through the exercise for chapter 6. The initial scenario on page 170 indicates that a positive test was returned for lizard flu. The question posed is what are the probability you have lizard flu given you tested positive. The first piece of data is that there is a 90% probability of lizard flu given a positive test and 9% given a negative test. Page 172 suggests that given a positive test, the probability of having lizard flu is less than 75%. After stepping through the methodology, adding new information about lizard flu (1% of the population has the lizard flu) there is the lizard flu diagnostic test that provides for "the probability that you have it, given that you tested negative" to calculate a 9% chance of having the lizard flu. Then on page 186 to change your base rate it says "9% of people who tested positively have lizard flu."

Can I get a little help understanding how the initial question of what is the actual probability you have lizard flu given you tested positive (p. 170) turned to the probability you have lizard flu given you tested negative (p. 180) concludes that 9% of people who tested positively have lizard flu (186)?

Thanks!

[blackermann]

I went through the Bayesian exercise again and all I can figure is that the formula at the bottom of p. 180 is worded incorrectly. It should read "The probability that you have it, given you tested positive" rather than "The probability that you have it, given you tested negative." For this formula, every instance of the word "negative" should be replaced with "positive."

We see the same formula again on page 187 but in this case it is correctly worded "The probablity that you have it given you tested negative."

[komyong]

I'm working through the exercise for chapter 6. The initial scenario on page 170 indicates that a positive test was returned for lizard flu. The question posed is what are the probability you have lizard flu given you tested positive. The first piece of data is that there is a 90% probability of lizard flu given a positive test and 9% given a negative test. Page 172 suggests that given a positive test, the probability of having lizard flu is less than 75%. After stepping through the methodology, adding new information about lizard flu (1% of the population has the lizard flu) there is the lizard flu diagnostic test that provides for "the probability that you have it, given that you tested negative" to calculate a 9% chance of having the lizard flu. Then on page 186 to change your base rate it says "9% of people who tested positively have lizard flu."

Can I get a little help understanding how the initial question of what is the actual probability you have lizard flu given you tested positive (p. 170) turned to the probability you have lizard flu given you tested negative (p. 180) concludes that 9% of people who tested positively have lizard flu (186)?

Thanks!

[sonia1s2o]

Hello,

I too had a similar question. Any answers to it?

Thanks,
Sonia

[bowie101@email.com]

http://oreilly.com/c...n=9780596153946 <--errata for the book

Note from the Author or Editor:
There is an error here. The math is correct -- there is a 9% chance that you have lizard flu. But the equation itself is incorrectly worded. It should be: The probability that you have it given you tested positive = # people who have it and test positive / (# who have it and test positive + # who don't have it and test positive) The error should be fixed on both p 179 and 180.