Re: Experiments With Human Subjects

From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Tue Sep 24 1996 - 07:20:14 MDT


On Mon, 23 Sep 1996, Robin Hanson wrote:

> Chris Hibbert writes:
> >I think you're drawing a tighter line around what counts as an
> >experiment than you intended. ...
> >They can move telescopes around, make predictions for other celestial
> >objects and verify them, do the observations from spacecraft or
> >aircraft as well as ground-based observatories, etc. This seems to be
> >fully as controllable an experiment as you need to become convinced of
> >the predictive value of Einstein's conjecture.
>
> If there is any sensible distinction between observation and
> experiment, it is in one's degree of control over a system. Observers

I think there is none. Observation is part of the experiment. Afaik there
is nothing in the definition of the experimental setup requiring it to be
created by human experimentators. It has just to be well defined.

Exactly as you wrote, experimental setups & observation (measurement)
procedere are designed to reduce impacts of undesired variables, ideally
excluding them altogether.

In social sciences experiments are usually impossible. Even observations
of system's behavious, are difficult, since isolated systems either do
not exist or the entire range of relevant parameters is impossible to
obtain and/or to interprete.

> look, while experimenters change things (and look). Of course
> observers can look at things from different angles, at different
> times, at different frequencies, etc. And they can choose what to

Yes. This is a change of parameters. The universe is one great lab. By
choosing appropriate observation locations much can be learned.

> focus attention on. But if obervation vs. experiment means anything,
> this type of control does not imply its an experiment.

I think you are using an unconventional definition of an experiment.
 
'gene

> Robin D. Hanson hanson@hss.caltech.edu http://hss.caltech.edu/~hanson/



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:35:45 MST