Confidence

[cheheyr at DHW.STATE.ID.US]

Here's a practical question for those of you who have complained about
philosophical
discussions on TAXACOM, and any others who can help. Our laboratory is doing
taxonomy
on insects, fungi, and algae.  In addition, we oversee contracts with
private parties and
universities for our overflow work.

One of our largest clients has asked that we develop a method of quantifying
our confidence
in our determinations.  This is amazingly easy when one is using numerical
methods, but is
eluding us with Linnaean determinations. It is basically a data entry
problem.  They want a
short field, by which an analyst can quickly and reliably determine the
quality of the data.
Such a protocol, if we all work from the same keys, might also help to
smooth differences
between taxonomists.

Our thoughts are running to a scoring system.  For instance assigning
weighted numerical
values to such variables as closeness of fit to a written description, stage
of development,
wholeness of the specimen, and size of the series.  Scores would then be
added and a range
produced.

I realize that little or no objectivity will be injected into our system by
this endeavor,
however if some uniformity happens, it will be a good thing.  I would
appreciate any
comments or suggestions, positive or negative.  Even if your answer is
"Bhtlfsk and
Doodermeyer did this in 1650, you dolt!", at least we'll have a starting
place.



*****************************************************************
 Robert L. Chehey             Phone: (208) 334-2235 x252
Manager, Microbiology              Fax: (208) 334-2382
Department of Health and Welfare,
Bureau of Laboratories
2220 Old Penitentiary Road         E-Mail:
Boise, ID,  83712                  Cheheyr at dhw.state.id.us
*****************************************************************