"Ned misses an important point, that statistical models don't give you any
idea what to measure. They simply tell you how to do what you are planning
to do anyway in a way that might give statistically meaningful results. They
do not have any underlying natural structure."--Bill


This is in fact part of the point I was attempting to make in regards to
your mention of those doing naturalistic research. Such research does not
necessarily require, or in many cases allow, a rigorous articulation of what
is being asked.

This can be contrasted with researchers attempting to exploit a neighbor
statistician to analyze data collected in the absence of an initial
appreciation of how a question was being asked and how that question fits
into broader theory. 

I'd also think it goes without saying that the statistical model  you're
testing, that is the independent variables (i.e. the causal factors) and how
they relate to the dependent variables needs to be based on biology, an
understanding of the system you're interested in and theory. An absence of
such would certainly be characterized as bad science.


Ned Dochtermann




----- Original Message -----
From: "Ned Dochtermann" <[EMAIL PROTECTED]>
To: "'Bill Silvert'" <[EMAIL PROTECTED]>; <[email protected]>
Sent: Thursday, March 09, 2006 4:23 PM
Subject: RE: "Hamerstrom science" (deliberate non-use of statistical
analysis)


> Generally however those concerned, after the fact, about the rigor of 
> their statistics (or lack thereof) are not reporting naturalistic 
> observations but attempting to hammer their round data pegs into the 
> square holes of already established theory.
>
> If your concern is naturalistic observation, you don't have to have 
> too much concern about whether or not you've properly articulated (or 
> understand) the underlying statistical model you're testing.
>
> Ned Dochtermann

-----Original Message-----
From: Ecological Society of America: grants, jobs, news
[mailto:[EMAIL PROTECTED] On Behalf Of Bill Silvert
Sent: Thursday, March 09, 2006 4:29 AM
To: [email protected]
Subject: Re: "Hamerstrom science" (deliberate non-use of statistical
analysis)

Although Mike doesn't use the term, this is a nicely put statement of the
message that modellers have been trying to get across for eons, that one
should model a system before doing the field work in order to design the
experiments optimally. Too often I have had people approach me with masses
of data, but without the critical information that is needed to understand
the system.

On the other hand, if one only carries out field work to test pre-existing
ideas, how can you discover anything new? One of the greatest scientific
events of the past century was the discovery of ecosystems based on
chemosynthesis rather than photosynthesis, but this was just the result of
sending down a ROV and had nothing to do with hypothesis testing. And Darwin
did not set out to test evolution, he joined the Beagle as a field
naturalist and developed his theory from his observations. I suspect that
these and other major scientific developments would not pass the rigorous
tests of "correct science".

Bill Silvert

Reply via email to