Drexel dragonThe Math ForumDonate to the Math Forum



Search All of the Math Forum:

Views expressed in these public forums are not endorsed by Drexel University or The Math Forum.


Math Forum » Discussions » sci.math.* » sci.stat.math.independent

Topic: small confidence limit to demonstrate inability to reject H0?
Replies: 4   Last Post: Oct 13, 2013 4:09 PM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View   Messages: [ Previous | Next ]
Richard Ulrich

Posts: 2,862
Registered: 12/13/04
Re: small confidence limit to demonstrate inability to reject H0?
Posted: Oct 11, 2013 6:48 PM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply

On Fri, 11 Oct 2013 07:29:35 -0700 (PDT), paul.domaskis@gmail.com
wrote:

>On Thursday, October 10, 2013 7:08:03 PM UTC-4, Rich Ulrich wrote:
>
> <...some confirmations that small p-values do not demonstrate a
> lack difference, and that the use of large p-values may be due to
> historical lack of understanding, which may be improving...>
>

>> As a very general rule, p-levels are not good indicators of effect
>> size. Or, conversely, for lack-of-effect size. N is the huge
>> confounder, plus, it depends on how sensitive your test is for
>> finding an "effect."

>
>Understood. I'm just focusing on acceptance/rejection of H0.

>>
>> When you want to show similarity, you want to show a narrow CI on
>> some relevant difference, for some meaningful measure... not merely
>> an absence of "demonstrated" difference. Crappy measures can give
>> you apparent similarities.

>
>My understanding is that narrow CI can only show lack of evidence against H0, but can never show the evidence for H0.


As Herman says in a post that *seems* to reply to this
(although he tacked on to your other comment from about
the same time), it is an easy conventional comment that
"The null hypothesis is never true."

A narrow CI on the size of a difference gives us evidence
that the difference is not larger. Or, not large enough
to matter.

And I don't worry about the Bayeisian approach.

--
Rich Ulrich




Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© Drexel University 1994-2014. All Rights Reserved.
The Math Forum is a research and educational enterprise of the Drexel University School of Education.