Trying to determine the overall probability of two false results for the permutations of two binary tests

[Lauri Bee]

1. The first question I'm faced with is to determine the probably of all false results when taking two rapid covid tests (2 false positives, 2 false negatives, false negative/false positive, false positive/false negative) when True Positive rate is 84% and True Negative is 98.5%. The goal is to determine how likely it is to get two false tests.

2. A second question was how like it is to get at least one positive result (with the same rates).

I am not sure of the formula to determine the results. The best guess I've been able to come up with is

Question 1:
(FP*FP)+(FN*FP)+(FP*FN)+(FN*FN)
FP= False Positive
FN=False Negative

or

(.015*.015)+(.015*.16)+(.16*.015)+(.16*.16)=.028

2.8% doesn't seem right however, I think I'm doing something wrong

Question 2:
(TP*TP)+(TP*FN)+(FN*TP)+(FP*FP)+(FP*TN)+(TN*FP)
(.86*.86)+(.86*.16)+(.16*.86)+(.015*.015)+(.015*.985)+(.985*.015)= 1.010175

101.0175% doesn't seem right for the probability of getting at least one true result

Report this ad​

0 Answers​

order by active oldest votes
Answer Your Question

meta chat tour help blog privacy policy legal contact us cookie settings full site
​

 

[Steven G]

The 2nd answer must be wrong since the probability of any even must be between 0 and 1 inclusive. 1.010175 is not in that range.
In number 1, is it possible, using your method with different numbers, that you can get a result more than 1? If yes, then your method is flawed.

Have you heard of conditional probability?