Re: Fallacy of Hypothesis Testing

From: Richard Ulrich (Rich.Ulrich_at_comcast.net)
Date: 02/25/05 

Date: Fri, 25 Feb 2005 18:42:06 -0500

 - this example doesn't nail the flaw ....

On Fri, 25 Feb 2005 20:20:03 GMT, passerby <passerby@yahoo.com> wrote:

> We all know that the following argument is valid:
> If H is true then E is false.
> E is true.
> Hence, H is false.
>
> However, the following similar argument is INVALID though almost all
> hypothesis testings are based on it:
>
> If H is true then the probability that E is true is (very) low.
 *because* E (largely) contradicts it ...
> E is true.
>
> Hence, the probability that H is true is (very) low.
>
> Following is a counter example:
> Consider this argument:
> (P1) If A is working in US then the probability that A is working at New
> York is (very) low.
> (P2)A is working at New York.

But P2 does not contradict P1, at all...
> Hence,
> (C) the probability that A is working in US is (very) low.
>
> (P1) and (P2) are true but (C) is false (all people working at New York
> must be working in US, so the probability that A is working in US is 1
> instead of (very) low!).
> Thus the above-mentioned argument form is INVALID. 

-- 
Rich Ulrich, wpilib@Pitt.edu
http://www.pitt.edu/~wpilib/index.html

Relevant Pages

  • Re: Psychosama: The Scariest Film of All Time
    ... Ion says the probability that tomorrow ... The New York Lottery Commission ... spinning better than any of us, the normals, because ...
    (rec.gambling.lottery)
  • Re: Whats the difference between the two sentences? (Using would and may/might)
    ... He would have reached New York by now. ... That's one meaning, but it could be the answer to a hypothetical question ... expresses a slightly greater probability that he has reached New York. ... Mark Brader | "Perl is a minimalist language at heart. ...
    (alt.usage.english)
  • Re: Fallacy of Hypothesis Testing
    ... passerby wrote: ... > If H is true then the probability that E is true is low. ... > A is working at New York. ... > Thus the above-mentioned argument form is INVALID. ...
    (sci.stat.math)
  • Fallacy of Hypothesis Testing
    ... hypothesis testings are based on it: ... If H is true then the probability that E is true is low. ... York is low. ... Thus the above-mentioned argument form is INVALID. ...
    (sci.stat.math)
  • Re: A question on Shannon entropy
    ... Jean-Luc Cooke wrote: ... > - each symbol in X has probability of appearing equal to H ... and your calculation will be invalid in that case. ...
    (sci.crypt)