Notes from Pat Nahas

Client/statistician mutual understanding,
Documentation which is clear to the reader
Importance of data archives
Expert witness testimony

Date: Wed, 20 May 1998 12:30:43 -0600
From: "Pat Nahas"

Based on a first-pass review, I would like to offer the following
comments:

1. [After further correspondence, Pat Nahas' comments on deception
appear under that heading.]

2. Under B, I would add an item that addresses arriving at a clear,
mutually agreed upon, documented statement of work prior to the
commencement of the work. It should include what the
customer/client/funder/employer expects in terms of deliverables -
information/results /reporting formats, etc. It should define what the
researcher/consultant/supplier needs from the customer in order to
accomplish the work - information, software, data - as well as who the
provider will be. The statement of work should also address
uncertainties/risks and time schedules, as well as ownership of
intellectual property and publication rights. A good project management
approach can go a long way toward eliminating the misunderstandings and
confrontations that often challenge ethical behavior.

3. Some mention of the need to provide clear and accurate documentation,
in language that can be understood by the intended audience - that is
not subject to multiple interpretations - is needed. In particular,
differences in the meanings of terms often leads to misinterpretations.
And lack of clear explanations - particularly when the information
provided is being used for decision making by people who are not
statistically knowledgeable - is not ethical. For example, a report
stating a % confidence of nn% may be technically correct. However, the
provider should take into consideration the level of knowledge of the
audience and the expected use of the information and provide enough
additional information to enable wise decisions. 95% confidence may be
acceptable for deciding between to fertilizers, but not for a medical
screening result that could lead to misdiagnosis.

4. Although it probably does not fall specifically under "ethics," I
would like to suggest that, since we have broken out several
responsibilities, we include an additional one that addresses
specifically maintaining records/data for historical purposes, for
reuse, etc. There is an implicit assumption in your listed items that
this occurs - and in many instances, particularly with respect to
academic research and professional publications, this is the case.
However, within industry, I have not observed this to be so much the
case. Experimental data is often lost/left behind when, for example, an
engineer moves to another position and their computer/files is "purged."
Much learning is lost.

5. Lastly, in the Preamble, some mention of Expert Witness Testimony
seems in order under B. Statistics and Society.

As I mentioned, I had limited time to review the guidelines. In general,
they are an impressive and well-thought through piece of work.

Pat Nahas (Patricia Lechner Nahas)
Motorola - Semiconductor Products Sector
__________________________________________________

Each of these topics deserves additional discussion. Please comment.
Thanks to Pat Nahas for raising them.