is my professor wrong or am i just stubborn?!!!

[americanpierg]

Please help me out with your answer to this problem, it could very well mean the difference between passing the class and getting my transcript in with a MODERATELY decent gpa, or failing the class and totally screwing up the poor chance I already have. I have to get these darn summer scores in before I apply in these upcoming weeks.

Incidence of TB in the US is 5 cases per 10,000. A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person) and 1% of false negative (negative result for a sick person). Find the probability that a person who tested positively actually has TB.

PLEASE tell me why the answer is NOT 99 percent??!!!

---

Members don't see this ad.

 

[DrBowtie]

What does your prof say the answer is?

 

[americanpierg]

What does your prof say the answer is?

i dont remember exactly but it was VERY low, ie <5%, (or was it <5%), so i didnt even care to remember it because i thought at the time i was 100% right

 

[polarnut]

What does your prof say the answer is?

 

[americanpierg]

lol so we meet again mr nut

 

[polarnut]

how can it be 99% when only 5 out of 10,000 people actually have TB. that's 0.05% now factor in the 1% false postitive stat, your answer should be, indeed, less than 0.05%

 

[DrBowtie]

how can it be 99% when only 5 out of 10,000 people actually have TB. that's 0.05% now factor in the 1% false postitive stat, your answer should be, indeed, less than 0.05%

I think somewhere the question isn't the same or somone misinterpreted.

 

[polarnut]

I think somewhere the question isn't the same or somone misinterpreted.

dude i'll own that professor

 

[americanpierg]

I think somewhere the question isn't the same or somone misinterpreted.

Thats what i was trying to explain to the prof.

Incidence of TB in the US is 5 cases per 10,000. A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person) and 1% of false negative (negative result for a sick person). Find the probability that a person who tested positively actually has TB.

She AGREES that the pool of people being "surveyed" already tested positive. Therefore, theres two options, 99/100 people will be "YES" (has TB) or 1/100 will be "NO" (does not have TB). The 5/10,000 becomes irrevelent in the way the wording of the question is.

NOW if the question becomes, Find the probability that a person will test positively and have TB, then itd become a very small percentage, below 1 percent even

 

[DrBowtie]

Thats what i was trying to explain to the prof.

Incidence of TB in the US is 5 cases per 10,000. A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person) and 1% of false negative (negative result for a sick person). Find the probability that a person who tested positively actually has TB.

She AGREES that the pool of people being "surveyed" already tested positive. Therefore, theres two options, 99/100 people will be "YES" (has TB) or 1/100 will be "NO" (does not have TB). The 5/10,000 becomes irrevelent in the way the wording of the question is.

NOW if the question becomes, Find the probability that a person will test positively and have TB, then itd become a very small percentage, below 1 percent even

Well which is the question on the test?

 

[americanpierg]

Well which is the question on the test?

Find the probability that a person who tested positively actually has TB.

 

[abzirian]

I have to agree with your 99% - I don't see how incidence of TB and chance of a false negative is relative to the question. If the test is 99% accurate, then the answer should be straight-forwardly a 99% chance of a correct result.

 

[Hardbody]

Please help me out with your answer to this problem, it could very well mean the difference between passing the class and getting my transcript in with a MODERATELY decent gpa, or failing the class and totally screwing up the poor chance I already have. I have to get these darn summer scores in before I apply in these upcoming weeks.

Incidence of TB in the US is 5 cases per 10,000. A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person) and 1% of false negative (negative result for a sick person). Find the probability that a person who tested positively actually has TB.

PLEASE tell me why the answer is NOT 99 percent??!!!

I wanna flame you on this one for knocking all the "dumb CC students" but I won't, I will just explain the answer.

9,995x.01=99.95 false positives (healthy people)+ 5 true positives

5/104.5x100=4.78%

 

[Hardbody]

Please help me out with your answer to this problem, it could very well mean the difference between passing the class and getting my transcript in with a MODERATELY decent gpa, or failing the class and totally screwing up the poor chance I already have. I have to get these darn summer scores in before I apply in these upcoming weeks.

Incidence of TB in the US is 5 cases per 10,000. A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person) and 1% of false negative (negative result for a sick person). Find the probability that a person who tested positively actually has TB.

PLEASE tell me why the answer is NOT 99 percent??!!!

Approximately 104.5 people will test positive, but only 5 of those people that tested positive actually have TB. This is a very simple problem.

 

[GoJoeyMojo]

I wanna flame you on this one for knocking all the "dumb CC students" but I won't, I will just explain the answer.

9,995x.01=99.95 false positives (healthy people)+ 5 true positives

5/104.5x100=4.78%

99% of the positive tests are accurate. But if you're taking .01 (accounting for the false positive tests) and applying it to 9,995, wouldn't you be saying all 9,995 are positive tests? I don't know if I'm making much sense.

 

[americanpierg]

I wanna flame you on this one for knocking all the "dumb CC students" but I won't, I will just explain the answer.

9,995x.01=99.95 false positives (healthy people)+ 5 true positives

5/104.5x100=4.78%

theres something wrong in there...somewhere

 

[Hardbody]

99% of the positive tests are accurate. But if you're taking .01 (accounting for the inaccurate tests) and applying it to 9,995, wouldn't you be saying all 9,995 are positive tests. I don't know if I'm making much sense.

The question asks to look strictly at all positive tests. You can assume that all TB infected people will test postive (5 positive tests) and 1% of all healthy people will test positive (9,995x.01=99.5). This means that 99.5 healthy people out of 9,995 WILL test positive.

Now, the combined number of ALL positive tests are 104.5. Remember, the question asks, "Find the probability that a person who tested positively actually has TB." 5/104.5x100=4.78%. Please take some time to think about it before posting again.

 

[Hardbody]

theres something wrong in there...somewhere

There is nothing wrong. It is a simple question. You are reading too much into the question.

 

[americanpierg]

The question asks to look strictly at all positive tests. You can assume that all TB infected people will test postive (5 positive tests) and 1% of all healthy people will test positive (9,995x.01=99.5). This means that 99.5 healthy people out of 9,995 WILL test positive.

Now, the combined number of ALL positive tests are 104.5. Remember, the question asks, "Find the probability that a person who tested positively actually has TB." 5/104.5x100=4.78%. Please take some time to think about it before posting again.

LOL the order of steps makes sense...but the outcome doesnt! That would represent like 95.22% error.............

 

[GoJoeyMojo]

The question asks to look strictly at all positive tests. You can assume that all TB infected people will test postive (5 positive tests) and 1% of all healthy people will test positive (9,995x.01=99.5). This means that 99.5 healthy people out of 9,995 WILL test positive.

Now, the combined number of ALL positive tests are 104.5. Remember, the question asks, "Find the probability that a person who tested positively actually has TB." 5/104.5x100=4.78%. Please take some time to think about it before posting again.

Yes but you are saying that 1% of all healthy people will test positive. How do you infer that from the statement saying 99/100 positive tests are false positives? That isn't the same thing as saying 1/100 healthy people test positive. And what's the attitude? Jeez guy, be nice.

 

[Hardbody]

what's the attitude? Jeez guy, be nice.

Sorry if you took attitude out of my last sentence. It is more frustration than anything. I will say this in generally to anyone that does not agree with me on the answer. GET SOME SLEEP!!!!!! If you can't figure this out by now, I would be willing to bet you are just plain tired.

 

[Hardbody]

Yes but you are saying that 1% of all healthy people will test positive. How do you infer that from the statement saying 99/100 positive tests are false positives?

"A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person)"

 

[dat_student]

Please help me out with your answer to this problem, it could very well mean the difference between passing the class and getting my transcript in with a MODERATELY decent gpa, or failing the class and totally screwing up the poor chance I already have. I have to get these darn summer scores in before I apply in these upcoming weeks.

Incidence of TB in the US is 5 cases per 10,000. A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person) and 1% of false negative (negative result for a sick person). Find the probability that a person who tested positively actually has TB.

PLEASE tell me why the answer is NOT 99 percent??!!!

..........................----> 4.95 + (because 99% accurate)
........................./
...........---> 5 TB-----> .05 false negative
........../
10,000
.........\----> 9,995 No TB----> 99.95 false positive
..................................\
...................................----> 9,900.05 - (because 99% accurate)

......4.95
P = ------------- = 0.047187798 (i.e. 4.72%)
......(99.95+4.95)

 

[GoJoeyMojo]

"A test for TB is 99% accurate, that is it gives 1% of false positive (positive results for a healthy person)"

Okay. I see. I think I'm operating under the assumption that they meant 1% of false positives refers to 1% of all positive tests are false positives. You are operating under the assumption that 1% of false positives refers to 1% of tests result in false positives for healthy people.
Me: 1/100 positive tests are wrong
You: 1/100 healthy people test positive
After reading the wording, you're probably right. It's late. And besides, I ain't the sharpest marble in the bag.

 

[dat_student]

I wanna flame you on this one for knocking all the "dumb CC students" but I won't, I will just explain the answer.

9,995x.01=99.95 false positives (healthy people)+ 5 true positives

5/104.5x100=4.78%

"99.95 false positives (healthy people)+ 5 true positives" >> Wrong
It's 4.95/(99.95 + 4.95)

see above

 

[americanpierg]

Wrong
it's 99.95 + 4.95

see above

...lol dont be picky, it gives you roughly the same answer so i didnt complain about that, its the big picture that counts.

Can ANYONE validate the steps that HARDBODY did to receive that answer...it looks correct matematically but it doesnt make sense logically

 

[Hardbody]

"99.95 false positives (healthy people)+ 5 true positives" >> Wrong
it's 99.95 + 4.95

see above

For the sake of simplification I ignored false positives on the 5. Too many people on this thread seemed to have a hard enough time grasping the basics of the question, I did not want to complicate it anymore than it already was for them. If it was a MC question with very close answers I wouldn't have done that, but imo it is close enough. So yes, technically you are right.

 

[Hardbody]

I ain't the sharpest marble in the bag.

That makes two of us, but I am proud of it !

 

[dat_student]

...lol dont be picky, it gives you roughly the same answer so i didnt complain about that, its the big picture that counts.

Can ANYONE validate the steps that HARDBODY did to receive that answer...it looks correct matematically but it doesnt make sense logically

I guess you didn't see my 1st reply:

..........................----> 4.95 + (because 99% accurate)
........................./
...........---> 5 TB-----> .05 false negative
........../
10,000
.........\----> 9,995 No TB----> 99.95 false positive
..................................\
...................................----> 9,900.05 - (because 99% accurate)

......4.95
P = ------------- = 0.047187798 (i.e. 4.72%)
......(99.95+4.95)

For the sake of simplification I ignored false positives on the 5...

ok

 

[americanpierg]

1/6,000,000,000 will have the same DNA as that found on the victim's clothes. The DNA test guareentees 99.999% accuracy. 0.17% of all that tested positive will be the killer.... Why the hell are DNA tests taken so seriously in court?!!!

This thread changed my entire view on life.

 

[DrBowtie]

1/6,000,000,000 will have the same DNA as that found on the victim's clothes. The DNA test guareentees 99.999% accuracy. 0.17% of all that tested positive will be the killer.... Why the hell are DNA tests taken so seriously in court?!!!

This thread changed my entire view on life.

I'd say that beyond a reasonable doubt.

 

[Lefty7]

1/6,000,000,000 will have the same DNA as that found on the victim's clothes. The DNA test guareentees 99.999% accuracy. 0.17% of all that tested positive will be the killer.... Why the hell are DNA tests taken so seriously in court?!!!

This thread changed my entire view on life.

^
|
|

Don't go changing your life's philosophy just yet, especially over a few statistics. They should have taught you the first day of your stats class that numbers can say almost anything depending on how you calculate them. Advertisers love this because they can use "hard numbers" to validate their product by manipulating the data pool.

Now, regarding DNA testing... Do you really think the test pool is 6 billion? 0.17% is only valid if you test everyone on the planet. By the way, even if you wanted to test everyone to find a match, the DNA SSR's more than unique enough to provide it. It is a simple enough thing to change the assay to include a few more probes to up the accuracy even more. I suppose forensic scientists just thought that a mere 99.999% was good enough to hold weight in an identity arguement.

Seriously, if you're really concerned about DNA profiling accuracy, I would focus on the possibility of DNA contamination during sample collection amplified by PCR. The test itself is foolproof, you can't reasonably argue a match is not a match, and if you did, it is easy enough to retest to confirm results. You just have to make sure the sample DNA being matched really is the DNA of the killer.

I don't know who told you that only 0.17% of positive tests in court are really the killer, but you might want to think about re-educating them.

 

[drnpp]

5/10,000 X 99/100 = 0.0495 x 100 = 4.95%

thats just a rough guess from what remember how to do this kind of a problem

 

[ObviousGuy]

who the hell cares?

just skip the question, get your b+ and move on.
WTF?

4.95%

 

[organichemistry]

edited out

 

[drnpp]

The question is asking "Find the probability that a person who tested positively actually has TB" not "If a person tested positive what is the probablity that they actually have TB"

So you have to take the 5/10,000 into consideration and can not ignore it which will give approx 4.95%

 

[Hardbody]

5/10,000 X 99/100 = 0.0495 x 100 = 4.95%

thats just a rough guess from what remember how to do this kind of a problem

What is this? The fuzzy math method?

 

[Hardbody]

The question is asking "Find the probability that a person who tested positively actually has TB" not "If a person tested positive what is the probablity that they actually have TB"

These two statements mean exactly the same thing. That being said, nobody read the second statement, since it was not posted on this thread.