Re: DBT in audio - a protocol

On 13 Jan 2006 12:46:20 -0800, "andy" <andy19191@xxxxxxxxxx> wrote:

>Don Pearce wrote:
>> So I make the test sighted, and the subject scores 100%, saying of
>> course they sound different. How do I factor that out of the
>> experiment?
>
>If 50% of the time they were the same then the results are perfectly
>correlated with sight and perfectly uncorrelated with sound.
>
>If 100% of the time they were actually different then the experiment
>was badly designed and the bias cannot be determined.
>
I haven't quite understood what you are saying - so I will try and
phrase mine a different way. If I make such a test sighted, then it
won't matter what the sound does - or doesn't do - the subject will
report the sighted case every time. This is why the test must be
blind.

>> I repeat. Sighted tests are valueless.
>
>No. Biased experiments are more difficult and less accurate to work
>with but valid results can be extracted in many cases where handling
>the bias is part of the design of the experiment.
>
In this case it isn't simply a matter of bias making analysis tricky.
The bias makes analysis totally impossible, because the bias is all
that remains in the experiment.

>However, I was only making a suggestion and if you are not happy with
>it then please just forget it.

No, please try again. I really didn't understand how sighted bias
could be factored out of this situation.

d

Pearce Consulting
http://www.pearce.uk.com
.

Relevant Pages

  • Re: DBT in audio - a protocol
    ... >Many experiments are performed in the presence of bias. ... course they sound different. ... Pearce Consulting ...
    (uk.rec.audio)
  • Re: DBT in audio - a protocol
    ... Don Pearce wrote: ... > So I make the test sighted, and the subject scores 100%, saying of ... correlated with sight and perfectly uncorrelated with sound. ... was badly designed and the bias cannot be determined. ...
    (uk.rec.audio)
  • Re: Its amazing what you can find when you look.
    ... >>> will sound just their old gear. ... >>> Of course he'd say that he knows how unrelaible such listening is, ... >> bias, it's bias toward the budget item, not the luxury item. ...
    (rec.audio.opinion)
  • Re: Strange noise from my amplifier
    ... So...two different things unrelated to the new power tubes. ... harmonic-rich sound. ... That's where I would disagree with the 'only one bias point' comment. ... When I first used a DiMarzio Super Distortion, ...
    (alt.guitar.amps)
  • Re: Question about sighted testing
    ... >> Actual purchase decisions are influenced by factors other than sound ... Bias obviously takes its "toll" ... > on the validity of these judgments. ... > do with home audio on the consumer level. ...
    (rec.audio.opinion)