Good and bad data (from Mary)

[from Mary Powers 960904]

Hans Blom (960903)

     What do you mean by GOOD data? [etc., etc., blah blah blah]

Hans, I think you are being faux naif about what I said. By
"good data" I did not mean just data that would support the PCT
model - and I think you know that but chose to interpret it that
way simply to score points against PCT, as you are so fond of
doing.

By good data I meant, as I am quite sure you know, data taken
while one controls factors which are completely extraneous to the
experiment and which would introduce artifacts that would confuse
the picture. One of these factors in the sleep deprivation study
was the failure to let the subjects practice tracking tasks to
the point of consistent competence before the experimental runs.
Some subjects on some tasks were sufficiently practiced, so that
their performance on those tasks, before the experiment and
during it, could be legitimately compared. Good data. Other
subjects failed to learn some of the tasks before the experiment
began, or ever. They had not learned to control, so it was
impossible to measure whether or how much they were able to
maintain control under the experimental conditions. Bad data.

So some of the data was good, and some was very bad indeed. This
says something about the research design and procedures, but it
doesn't really say anything about PCT one way or the other. And
I can do without a lecture on good and bad science, data, and
models that implies that I don't know these things. That is
insulting.

Mary P.

[From Bruce Gregory (960904.1150 EDT)]

(Mary Powers 960904)

There is no bad data, only bad interpretations of data.

Some subjects on some tasks were sufficiently practiced, so that
their performance on those tasks, before the experiment and
during it, could be legitimately compared. Good data. Other
subjects failed to learn some of the tasks before the experiment
began, or ever. They had not learned to control, so it was
impossible to measure whether or how much they were able to
maintain control under the experimental conditions. Bad data.

One person's good data is another person's bad data. If you had
wanted to demonstrate that subjects who cannot control to begin
with, do not gain the ability to control as a result of sleep
deprivation, the data you label bad would be labelled good. It
seems that much of the data in the experiment could not be used
to test PCT, so an opportunity was missed. That indeed was
unfortunate.

Best,

Bruce

[Hans Blom, 960909]

(Mary Powers 960904)

What do you mean by GOOD data? [etc., etc., blah blah blah]

So some of the data was good, and some was very bad indeed.
This says something about the research design and procedures,
but it doesn't really say anything about PCT one way or the
other. And I can do without a lecture on good and bad science,
data, and models that implies that I don't know these things.
That is insulting.

Mary, you put me in a difficult position. Let me start my reply with
a quote from a recent Wired magazine:

Among the professionals, philosophical argument is a martial art.
Nervous spectators pull their beer bottles out of the way in case a
backhand logic chop should sweep them off the table. This is
personal.

The young man defending himself in the Empire Bar is dressed like a
jobbing rock star: t-shirt, tight jeans around skinny legs, great
curly masses of brown hair reaching down his back. But he moves like
a fighter; every point launched against him is blocked, deftly, with
his outspread palms.

His opponent is older, still wiry, with black hair and moustache.
The more the young man blocks, the harder the older man's arms throw
new points. Soon he is arguing from his shoulders, like a boxer.
Finally, he seizes a beer bottle and thrusts it in front of his
opponent. "Look," he shouts. "There's only one thing I want to know.
Do you think this beer bottle has consciousness?"

There's a pause. "Well, it might have," says David Chalmers. A
ripple of appreciative relaxation runs round the audience. The bout
is over. One of the spectators takes the bottle from Bruce Mangan's
hand and carefully tears a strip from the label. He waves it in
front of Chalmers. "So what happens now?" he asks. "Where has the
consciousness gone in the paper? Has the label got its own little
consciousness?"

It's a game. It is also extremely serious.

So far this quote from an article published in August 1996 UK edition
of Wired Magazine. Now suppose that Chalmers had answered "This is
insulting. You're just making fun of my theory, as you so often have
in the past". That would worry me. It would mean to me that he could
not discriminate between an attack on himself and thus would like to
stop a critical investigation of his theory. That would worry me,
because I would consider it an attempt to stop a lawful scientific
inquiry by inappropriate means. Few people, I think, want to hurt
others' feelings. I had no idea that my remarks would hurt yours. An
"unintended side effect", a PCTer might say. But feelings are not the
issue here. An appropriate reply to the above hypothetical emotional
outburst would be, I think: "No, not at all. What I do is I'm
investigating the ultimate implications of your theory." Chalmers, by
the way, is far from crazy. It is just that his definition of
consciousness is a bit idiosyncratic. But why wouldn't he be allowed
to express his point of view? Why should he have to be afraid that
his view hurts others, who have a different view? Our discussions
here are mild, compared to those in some other circles. I'm sorry
that you feel insulted, but that will not stop me from exposing what
I think are inconsistencies, inadequacies, errors or even an
unscientific feel in a theory, just as you are welcome to expose what
you experience as errors in mine. Not in order to disqualify, but in
order to cut off unfruitful branches in an extremely extensive search
tree. And maybe even in order to show new directions to explore.
Please let us not stop any attempts of discovering the relative
merits of different theories because we feel insulted when our pet
theory is treated less than reverently. Please let us not declare
certain topics taboo because they have an emtional impact.

Maybe we ought to take a time-out, once in a while, and look at our
theories with a helicopter view. If we can go up that level ;-). The
article continues:

The crowd around Chalmers and Mangan at the Empire Bar is part of a
congregation of neuroscientists, philosophers, psychologists,
quantum physicists and AI gurus that has come in pilgrimage to
Tucson for a conference called Towards a Science of Consciousness.
There are almost a thousand people attending, and each participant
worth talking to has at least half a dozen good theories, mostly
incompatible. What holds all the theories and disciplines together
is simple. These people think that the nature of consciousness is
the most exciting intellectual frontier in the world today. It is
here that science seems to be closing in on the essence of what
makes us human.

That's the spirit in which I would like to see our discussions
continued. It is, in my opinion, far too early to have a single,
convincing, consistent theory of why humans are as they are. We're
still very much in the very first discovery phase. We have no
solutions, yet. We only have suggestions for what to explore next.
What holds us together here is simple, I think: We, too, think that
we investigate the most exciting intellectual frontier in the world
today.

Greetings,

Hans