Re: Official Trek XI Website Has Teaser Trailer

"Jaxtraw" <jax@[EMAIL PROTECTED]
> writes:

> The problem occurs when reality cannot effectively bite. 
> For instance, suppose somebody uses a bad statistical method and
> apparently eating tuna fish increases the risk of bowel cancer. 
> So, you stop eating tuna fish. You don't get bowel cancer.
> Does this support the hypothesis? Not really, no.  Perhaps you stop
> eating tuna, and get bowel cancer anyway. That doesn't falsify the
> hypothesis either. Neither do people who continue eating tuna and
> don't get bowel cancer.
>
> Well, it shoudl be scientifically testable.
> Follow some people for 40 years and note their tuna intake 
> and bowel cancer occurrence. The problem with that is, 
> firstly it takes 40 years, and secondly the association
> is sure to be so weak that it could be chance anyway.
> In a pragmatic sense it's effectively impossible to test such
> supposedly scientific assertions.

Your proposed study, which isn't looking at much information,
would indeed result in weak associations and be useless.
Just because you can design a poor study or bogus experiment
doesn't mean that good ones can't be designed and executed,
and result in good data and results leading to useful theories
that can be put into practice.

http://www.framinghamheartstudy.org/