False positive/negative

[Siddhant]

In the 2011 paper, Q7.

Aren't type 2 errors false negatives? The answer key says otherwise.

[NM]

The question says that, H0: Patient is healthy.

Now the doctor wants to minimize the chance of a test result showing patient being healthy when they are actually sick. So, HA: Patient is not healthy (When they are actually healthy).

Type 2 error: Accepting wrong Null hypothesis; which in this case would be accepting that patient is healthy when he is sick. So the result that is given here is false positive. (Results being positive i.e. indicating he is healthy).

[ashutosh]

Thats my doubt too.

Even I think it should be false negative.

[The Villain]

Our null hypothesis is H0=Patient is healthy and is found healthy
Alternate hypothesis is H1=Patient is sick and is found healthy.


So clearly its a type 2 error(false positive)

[mrittik]

Type II error is false negative error-thats right...but here qsn asked in some diff way...surely it is Type II error...why we tell it false negative? becoz the significant value is outside the bell-curve(as assumed normal distribution) so the right value>the hypothetical value....the hypothetical value-right value=>which is a -ve term....now we concentrate in this problem....here minimizing the problem of accepting the false null hypothesis which is healthy....so here right value>(the hypothetical value)1 & for the doctor right value>(the hypothetical value)2....which is solely for the purpose of minimizing the Type II error....where (the hypothetical value)1<(the hypothetical value)2,  ((the hypothetical value)2-right value) - ((the hypothetical value)1-right value)....which is +ve term...thats why this Type II error is false +ve for the purpose of minimizing the error.....plz check & comment...kudos

[Siddhant]

Generally, the null hypothesis is that the patient is healthy/ product is not faulty etc. So the question does follow the normal pattern, so should there be a different answer? Like for a medical report, a positive result is generally bad news.

[mrittik]

I can't understand ur intention....u just pointed the medical report rightway....+ve report obviously is a bad/threatening outcome but here the situation is different. A man is assured by the Report that he is healthy but actually he is not. What is the very common feature in rural treatment where everyone is seemingly healthy bearing disease....here the qsn is to find out the disease bearing factors where it turns out from the ignorance...then if he is bearing a disease then its result should come out as you mentioned....+ve result is bad obviously but bearing disease with ignorance more dangerous...I think so....what do u think?

[Homer Simpson]

Type II error is referred to as false negative. Check wikipedia for this:-

http://en.wikipedia.org/wiki/Type_I_and_type_II_errors

Secondly, one cant apply doctor-patient analogy for false positives and negatives!

here Ho = patient is healthy
       Ha = not healthy

Doctor wants to see if patient comes out healthy but actually isnt - meaning we accept a false null - this null is an average figure and sometimes it can be miscalculated/wrongly presented by medical researchers. the usage of "positive" could also mean a favorable medical test result - no disease or being "positive" for a disease! so we cant use that as a basis.

[mrittik]

great Tsuki....thats the underlying concept...Whatever Siddhant figure out I'm not convince at all....ur approach broadens my view too....gracias amiga