Attractiveness Multiplier, who cares?
Posted: Thu Apr 19, 2001 2:50 pm
I think it is a commentary on the state of our profession that this subject
has generated so much discussion and heat, almost to the exclusion of any
other topics on the over the past week or so.
Jay asked "does it matter"? Id go further and say "who cares"?
You may formulate variables like attractiveness in different ways, the
ultimate validity test being that it makes sense to the manager who is going
to make a decision based on the interpretation. His view is the only valid
view at the end of the day, and the proper role of the SD practitioner is to
help him to express and modify his mental models - not to present him with
answers. It is of course valuable to have different mathematical
formulations in the tool bag and to understand the ways in which they affect
model outputs - but most managers just need to understand that if they do
something, then the likely effect will be roughly positive or negative. And
if they believe that clever mathematical formulations will help them predict
quantitative outcomes, then more fool them. Anyway, as various
correspondants have pointed out, there is a wealth of existing technical
literature on this subject.
If there is one thing that this whole discussion highlights, however, it is
the futility of building large black box models that incorporate many (or
even a few) of these relationships. Far more productive and useful to use
small structural SD models as a focus for discussing the "position" and
"nature" of non-linear variables, so that "less" or "more" feels more
comfortable to a manager faced with a challenging decision. Seen this way,
the most potent skills of a good SD practitioner are facilitative rather
than technical.
Richard Stevenson
From: "Richard" <richard@cognitus.co.uk>
has generated so much discussion and heat, almost to the exclusion of any
other topics on the over the past week or so.
Jay asked "does it matter"? Id go further and say "who cares"?
You may formulate variables like attractiveness in different ways, the
ultimate validity test being that it makes sense to the manager who is going
to make a decision based on the interpretation. His view is the only valid
view at the end of the day, and the proper role of the SD practitioner is to
help him to express and modify his mental models - not to present him with
answers. It is of course valuable to have different mathematical
formulations in the tool bag and to understand the ways in which they affect
model outputs - but most managers just need to understand that if they do
something, then the likely effect will be roughly positive or negative. And
if they believe that clever mathematical formulations will help them predict
quantitative outcomes, then more fool them. Anyway, as various
correspondants have pointed out, there is a wealth of existing technical
literature on this subject.
If there is one thing that this whole discussion highlights, however, it is
the futility of building large black box models that incorporate many (or
even a few) of these relationships. Far more productive and useful to use
small structural SD models as a focus for discussing the "position" and
"nature" of non-linear variables, so that "less" or "more" feels more
comfortable to a manager faced with a challenging decision. Seen this way,
the most potent skills of a good SD practitioner are facilitative rather
than technical.
Richard Stevenson
From: "Richard" <richard@cognitus.co.uk>