Subject: Ethical Isues in Statistical Software: the "make or buy" decision

Following message was posted by Chuck LoPresti on August 27, 1998 at 14:41:43:


Ethical Issues in Statistical Software

The "make or buy" decision

Mon, 18 May 1998 12:48:38 -0700
d38716@gaviota.pnl.gov (Chuck LoPresti)

What is the position of the Ethical Guidelines with regard to using
well-known commercial or public domain software packages versus writing
one's own statistical application packages? This consideration moves us
into the area of software testing and reliability. For example; is it
ethical to use an untested package when a tested package is available?
Or, would a part of ethical practice be to discuss the statistical
package used for an analysis? Sometimes the analyst has to write a
program because nothing else that meets the need is readily available.
How should this situation be handled? Would the Committee agree that
this is a matter best left to individual good faith judgement? Should
it be mentioned as a concern in the Ethical Guidelines? [Chuck LoPresti
also commented on deception; those remarks are quoted under that
heading.]

Otherwise, these seem to me to be excellent guidelines, and I commend
ASA's Committee on Professional Ethics. And thank you for providing us
the opportunity to comment on the Guidelines.

Charles Lo Presti, Environmental and Health Sciences Division, Pacific
Northwest National Laboratory
___________________________________________________________
All comments are referred to the Committee on Professional Ethics. Here
is a preliminary response from Dr. Gardenier - There are a number of
issues surrounding the make or buy decision, including numerous
contextual variables. For example, using established software and
methods may receive higher weight in legal cases, where adherence to
precedent is usually considered a virtue. Conversely, originality may
be more desirable in a journal article. Pending better insight from
additional comments, I do not see how the Guidelines can address this
problem without a much more extended treatment than other items
currently in the draft. In general, I suspect that complex items should
be left to the maturity, judgment, and personal integrity of
individuals.