[JN] Blind testing issues

Thorsten Loesch thorsten.loesch at gmail.com
Tue Aug 25 20:12:14 CDT 2015


Hi,

> Be sure to control for all of the biases in your experimental design as
> well. The sleight-of-hand of presumed the "objectivity" of the experiment
> itself being one, among many others.

Of course. Those are the first to deal with actually.

> You see, "science" and the "experiment" and everything you bring to it are
> contingent cultural creations also.

Sure, every experiment, every theory has limits. Knowing them is essential.

> We made all of this up...it isn't somehow there independent of human
minds.

You just re-invented Hume's contention.

>> And you know what, I could not care if one of my subjects had a penchant
>> for buggery, was in fact a girl who likes to munch the carpet and so on.
>
> I would say that in audio the former is far more likely than the latter.

And you might even want to test if an Audiophile with a penchant for
buggery as bottom prefers different Amp's and Speakers (maybe he likes BBC
LS3/5A and British 25 Watt Class Amps) from one who is top (maybe that one
likes Altec 604's driven by SE Tubes) and one who is straight.

>> In information theory it is called context. The experiment generates
data.
>> Place data into a context, it becomes information.
>
> In anthropology, linguistics, semiotics, and so on, context is far more
> encompassing, including language, participants, cultural assumptions in
> play, and on and on.

All of that is context. You need to decide where to draw the line and what
to include/exclude to avoid paralyses by over-analysis.

> Information science al la Cherry et al has a strong behavioral bent, as
> does the "management science" also propounded at Bell Labs/Western
> Electric. This was the kind of science that was popular mid-century
> because like your naked sales attitude, they wanted _results_ and
> _control_ and if that involved creating a framework where people were
> basically lab rats and even the scientists were basically calculating
> machines, that's cool.

This style of science yielded some amazing results. In other words "The
Planes landed".

I would not call trying to understand peoples preferences and giving them
what they prefer anti-humanistic BTW.

> This is where the ABX dudes learned their science.

Not really, what they have is a caricature, a severe distortion. And no
"The Planes don't land".

This is precisely the difference between science and a cargo cult...

Ciao T

On 26 August 2015 at 05:09, Joe Roberts <jroberts at prismnet.com> wrote:

>
> >> It is impossible to control for biases because we don't even always know
> >> what they all are, but this is no reason not to try.
> >
> > Of course, but is possible to control against specific biases, knows
> ahead
> > of the experiment to exist.
> >
>
> Be sure to control for all of the biases in your experimental design as
> well. The sleight-of-hand of presumed the "objectivity" of the experiment
> itself being one, among many others.
>
> You see, "science" and the "experiment" and everything you bring to it are
> contingent cultural creations also.
>
> The gesture to move science somehow out of that net is a cultural trick,
> along the lines of
>
> "God answers all prayers"
>
> "WTF, I prayed for a week solid that my amp would magically fix itself."
>
> "God answered...but he said NO"
>
> Science declares itself non-subjective but that is itself a subjective,
> cultural belief that rather questionable. A few instruments and math help
> the illusion but they are all cultural creations also, and the terms
> employed and goals are culturally sanctioned.
>
> We made all of this up...it isn't somehow there independent of human minds.
>
> Science is part of culture.
>
>
> > And you know what, I could not care if one of my subjects had a penchant
> > for buggery, was in fact a girl who likes to munch the carpet and so on.
> >
>
> I would say that in audio the former is far more likely than the latter.
>
>
>
> > In information theory it is called context. The experiment generates
> data.
> > Place data into a context, it becomes information.
> >
>
>
> In anthropology, linguistics, semiotics, and so on, context is far more
> encompassing, including language, participants, cultural assumptions in
> play, and on and on. The experiment and all fo the assumptiions that went
> into it are part of the context. This context defines what is being done
> and what people see their roles in it to be. It is multi level and
> mindboggling in the complexity it introduces but there is no meaning
> without this sociocultural context, ever.
>
> Information science al la Cherry et al has a strong behavioral bent, as
> does the "management science" also propounded at Bell Labs/Western
> Electric. This was the kind of science that was popular mid-century
> because like your naked sales attitude, they wanted _results_ and
> _control_ and if that involved creating a framework where people were
> basically lab rats and even the scientists were basically calculating
> machines, that's cool.
>
> This is where the ABX dudes learned their science. By the mid 1930s,
> philosophers of science who invented all of this logico-positivistic
> hogwash had already given up on it. Check out "Vienna Circle of
> Positivism" and have a nice Austrian lager for me while you are at it. I
> prefer Stiegl.
>
> By this era, however, the weight of thinking in human sciences and
> philosophy and even advanced physics, was way beyond that reductionist and
> anti-humanistic stance.
>
> JR
> "The Revolution will not be Digitized"
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.soundlist.org/pipermail/sound/attachments/20150826/536ae270/attachment.html>


More information about the Sound mailing list