Drifting reference levels

[From Bill Powers (930106.1115)]

Avery Andrews (930106.1856) --

a) conventional psychologists aren't interested explaining

b) their quantitative standards for micro-phenomena are too
    low to prevent them from getting buried in trash.

I have a suspicion that this is true of the institutional
structure of psychology, though I suspect that (a) is false for
many individual psychologists, and even more for people on the
margins of psychology, such as AI-ers, broad minded linguists,

The only psychologists I know of of whom these statements aren't
true are those who have discovered that there are behavioral
experiments, however simple, that will generate reproducible and
accurate results for each indivual. Once having had such an
experience, such psychologists are no longer able to go back to
the way experimental psychology has been done in the past.


RE: picking up finger, moving it, putting it back on spot with
eyes closed.

My conjecture is that there is some component of the assortment
of control systems involved here that is being pushed to its
limits of resolution, kinesthetic memory being one obvious
possibility, but there are others.

That's a start. I would want to get some instrumentation on this
before settling for a statistical summary of the results. For
example, is there any systematic drift in the kinesthetic
reference level that can be related to the movements in the time
before the finger comes down again? If I apply random (but known)
disturbances in x and y while the finger is off the paper, can I
influence the final landing position? Does the reference signal
drift with time even if the finger doesn't move while it's in the

It seems to me that you could do this experiment with a mouse,
and get some real data on it.

( Moving back a little...)

so the only useful description of the results is a statistical

That would be giving up before you even start.