Re: [Harp-L] comb test

[Jonathan Ross <jross38@xxxxxxxxxxx>]

> > If test after test (imperfect as they may be) fails to show that
> > people can hear a difference due to comb materials, that is not
> > insignificant, as you seem to be saying. '
>
> The whole point of the null hypothesis is that it it doesn't say
> anything, that's what makes it null. "I aceccept the null
> hypothosis" mean exactly the same as "I haven't the foggiest
> idea".

Let's be specific.  In this case what the tests show is that people  
could not identify a timbral difference due to comb material.  I  
certainly agree that doesn't prove the null hypothesis, but it does  
show that under controlled circumstances people could not hear a  
difference.

> If test after test fails to prove anything at all then that dosn't
> give you any scientific grounds for drawing any conclusions,

There are two separate issues here: the scientific study of phenomena  
and the practical usage of data.  Of course you cannot prove that  
comb materials don't effect timbre (I never said you could), you can  
however show that people fail to hear a difference in repeated  
tests.  The practical and real-world usage of that is a different  
matter than the purely scientific question (and if I conflated the  
two I apologize).  In real world terms we can proceed as if comb  
material does not effect the timbre because it has not been shown to  
do so.



JR Ross