Hi Ken,
Thanks for your message.
I feel you are mistaken - perhaps we see things differently.
You said ' This is a case of sophisticated but tractable mathematical
modeling of well-understood SOCIAL behaviors.'
The situation is clearly socio-technical rather than purely social. It's
focus is influencing people voting. This involves social issues (it is
humans with all their influenced behaviours doing the voting) and it is
technical ( influence occurs through mass communication technology, travel
by vehicular technology, in the built environment, influenced by politicians
using the technologies of the campaign trial etc etc).
You said ' it doesn't fit any of the seven socio-technical systems' because
it use mathematical modelling resembling data mining rather than predictive
socio-technical modelling and could be done with enough staff doing enough
calculations.
There seems to be some confusion on your part. They were classes of
socio-technical problems and iIn the seven categories (see below), I
distinguished between:
A. 'Being able to predict the behaviour of outcomes in an individual's mind
(e.g by thinking, or intuition) or even a group of individuals in
discussion' , and,
B. 'Being able to predict the behaviour of outcomes using mathematical
methods' (i.e. mathematical no 'using computers')
Calculations are a mathematical method not something done in mind. Hence,
using your words, the example I gave of the use of data modelling by
Obama's team fits potentially into categories 3, 4, 5, and 6 of types of
socio-technical problem
The Obama team stated this mathematical modelling approach was used in
preference to using the human thinking of 'political expertise' . This
preference for mathematical modelling over human thinking and intuition
suggests the situation is not in Category 3 and potentially is not in
Category 4. Their choice to use mathematical modelling rather than the
human thinking of 'political expertise' means it is not in Category 6.
By implication, therefore it is most likely a socio-technical problem of
type Category 5.
Further, experience of these kinds of data is they involve both
interrelationships between factors acting on decisions made by subgroups
and time-dependent feedback loops with delays. For example, messages
intended to encourage female voters would be more effective in some areas
different times and at the same time might discourage, e.g. misogynist
voters. Messages influence discussions in the media which have other
feedback effects including influencing workplace discussions etc.
Decades ago this kind of data would be simply seen as counting the numbers
in particular categories. My feeling reading your post is that this is the
way you are thinking of this situation.
The world has moved on. As the article quotes those creating Obama's models,
this kind of mathematical modelling and analysis now involves creating
dynamic models, initially calibrated against large datasets but once
ratified can be used for predictions on the basis of a small number of data
points. The approach, as I have described in earlier posts, is to create a
model and test it, and then run scenarios against it to identify better
options. A further development will be to create meta-modelling that will
automatically identify the better solutions - same as for most areas of
design.
If as implied by the data and the article, the mathematical modelling
involves complex multifactor feedback loops predicting dynamic behaviour of
outcomes then I suggest on the basis of empirical evidence that this
prediction is beyond the ability of human to do 'in mind' by thinking ,
feeling or intuition. Hence, the problem is clearly in socio-technical
category 5.
If you think differently, I'd love to hear the reasons.
Warm regards,
Terry
--
Dr Terence Love
Tel: +61 (0)4 3497 5848
Fax:+61 (0)8 9305 7629
[log in to unmask]
--
The seven categories I suggested in the earlier post were:
1. Easy - not much thinking required to predict behaviour of outcomes and
identify solutions. Often feeling or intuition is successful enough.
2. More difficult - more thinking required and some benefits are found in
using expertise of multiple people to predict behaviour of outcomes and
identify solutions
3. Very difficult type A - problems that have many different aspects but
intrinsically it is possible using thinking and expertise of multiple
people to predict the behaviour of outcomes and identify solutions, perhaps
with some assistance from visual or methodological tools
4. Very difficult type B - problems that have many different aspects and
are intrinsically IMPOSSIBLE to predict the behaviour of outcomes and
identify solutions by using thinking and expertise of multiple people
5. Very difficult type C - problems that have many different aspects and
are intrinsically IMPOSSIBLE to predict the behaviour of outcomes and
identify solutions by using thinking and expertise of multiple people, but
POSSIBLE to predict the behaviour of outcomes using mathematical (or in
some cases analogical) modelling methods.
6. Very difficult type D - problems that have many different aspects and are
intrinsically IMPOSSIBLE to predict the behaviour of outcomes and identify
solutions by using thinking and expertise of multiple people, and POSSIBLE
to predict the behaviour of outcomes using mathematical (or in some cases
analogical) modelling methods, BUT there is an willingness or lack of
expertise to do so, and instead an attempt is made to guess outcomes or
attempt to persuade people to choose between possible interventions without
them having adequate knowledge. The outcome prognosis for this approach is
usually poor.
7. Impossible - problems for which it is INTRINSICALLY IMPOSSIBLE to predict
the behaviour of the outcomes of the situation and changes in outcome
behaviours due to any interventions in that situation. An example might be
the behaviour of a perpetual motion machine.
-----Original Message-----
From: PhD-Design - This list is for discussion of PhD studies and related
research in Design [mailto:[log in to unmask]] On Behalf Of Ken
Friedman
Sent: Saturday, 10 November 2012 3:17 PM
To: Dr Terence Love
Subject: Re: Bridge papers - reply to Don part 1
Dear Terry,
Thanks for the example, but it doesn't fit any of the seven socio-technical
systems in your earlier post.
This is a case of sophisticated but tractable mathematical modeling of
well-understood SOCIAL behaviors. This kind of mathematical modeling yields
reasonably accurate predictions based on well understood methods of social
measurement enhanced by a technical system.
This is not an example that predicts how a socio-technical system behaves.
This is a case is which modelers use a socio-technical system to predict how
a social system behaves and will behave based on accurate modeling of
current preferences and behaviors.
The reason that computers are valuable in this process is the large numbers
involved and the multiple potential parameters. This resembles data mining
rather than predictive socio-technical modeling. It yields predictions, but
there's nothing here that James Farley couldn't have done for Franklin
Delano Roosevelt if he had enough poll-takers and staff members to do the
calculations based on current models with today's highly diverse population
of 311,000,000 rather than the far more homogenous American population of
128,000,000 that Roosevelt governed in 1936.
If Obama had had someone who served him as skillfully as Farley served
Roosevelt, I suspect that Obama would have had a more successful first term
despite the opposition, amore comfortable margin in the popular vote, and
enough seats in the House to ensure completing his legislative agenda in the
second term.
This is social survey modeling for politics, not quantitative predictive
modeling of socio-technical systems. I can't see how this supports your
argument.
Yours,
Ken
-----------------------------------------------------------------
PhD-Design mailing list <[log in to unmask]>
Discussion of PhD studies and related research in Design
Subscribe or Unsubscribe at https://www.jiscmail.ac.uk/phd-design
-----------------------------------------------------------------
|