> >But for statistics as a social science would recognise that
> >real world data comprises artefacts that are the products of human
> >activities and the way we think about them.
>
> As soon as you leave the realm of theoretical Statistics (or
> mathematics)
> and computer-simulated data sets, one always run into
> artefacts and other
> data-related difficulties. I have spent a lot of time in
> analysing satellite
> images and you always get the same problems as in the social
> sciences. Why
> then a social science statistics?

Statistics as a social science would not see artefacts as a 'data-related
difficulty' but as part of the core of its subject matter.

Are the categorisations of statistics real?    In the case of statistics
related to human behaviour they are real because people believe them to be
real.  The general point is well established in the social science.   Cooley
wrote (Social Organisation,1909) that 'if people believe a thing to be true,
then it is true for them in its consequence'.

The question of whether statistical categorisations are real is a major
theme of Desrosieres history of statistical reasoning The_Politics_of
_Large_Numbers, 1998.   (reviewed in Radical Statistic 73).    I quote his
conclusions:

.. conventions defining objects really do give rise to realities, in as much
as these objects resist tests and other efforts to undo them.   This
principle of reality affords an exit from the dead-ended epistemological
opposition between these two complementary and complicitous enemies, the
realist and the relativist.   It does not deny the reality of things once
numerous persons refer to them to guide and coordinate their actions (p
337).

But statistics as a social science would not accept statistical
categorisations uncritically even when the categorisations are widely
accepted.  A classic case in social science is statistics of suicide.   But
any area of deviance - like criminal statistics and homelessness - is
properly subject to continued debate about statistical categorisations.
Statistics as a social science would extend this debate to other areas -
such as economic statistics.

> What does systematic production of evidence means? Is it
> something to do
> with causality or looking for 'truth'?

The production of statistics by the social scientist is expected to be
geared to such purposes.   But we all depend upon statistics that are the
by-product of administrative processes.   Such statistics are systematically
produced but are not geared to causality or 'truth' in any wide sense.
Statistics as a social science should be concerned with evaluating such
administrative statistics and complementing them with investigation of what
is left out or what is misleading.    The criteria in mind, as with the
social sciences generally, would be society as a 'client'.

Ray Thomas, Social Sciences, Open University
Tel: 01908 679081 Fax 01908 550401
Email: [log in to unmask]
35 Passmore, Milton Keynes MK6 3DY