Ray Thomas wrote, inter alia:
But I didn't find anything in [Coomarem's] bibliography that addressed the
difference
between statistics of natural populations (often defined as the subject of
statistical method) and statistics of human activities. The recognition
of statistics of human activities as social or organisational constructions
identifies their different character. But does anyone know if any of
these writings, or any other writings, that directly address the
relationship of this difference to statistical method?
Ray, would you say more about what, in your view, follows from this
distinction? Is it that human activities, being presumably purposive in a
way that one does not ascribe to the behaviour of other natural objects, are
not amenable to statistical summary and inference?
I would say that what takes human activity outside the ambit of statistics
is its conscious co-ordination (on the rare occasion when it has this
characteristic).
So long as human activity is atomistic in its organisation (even though the
individual actors are purposive) then it falls to be analysed by the methods
of statistical mechanics, whose results depend only on the particulate
nature of the entities and need no assumptions, other than particulateness,
about their nature or action.
A persuasive argument for this is set out in chapter 2 of Emmanuel Farjoun
and Moshe Machover (1983) "Laws of Chaos" (Verso). I've recently learned
that Edgeworth also made the same point -- but only very tangentially, and
he seems to have misapplied the metaphor...
Julian Wells
AU Business School
The Open University
Walton Hall
Milton Keynes
MK IAA
United Kingdom
+44 1908 654658
|