Terry asked an important question, one I am unable to answer. But the
reasons i can't answer might be informative.
On Sun, Sep 10, 2017 at 6:03 AM, Terence Love <[log in to unmask]> wrote:
> At some stage, this move towards the human-centred side of design will
> need change to gain the benefits of the human-related issues being
> analysed scientifically most likely via formal mathematical modelling and
> quantification. That is, a move back again towards more
> mathematically-based design theories and research
Why I don't think mathematical models of human behavior will ever be
useful.
And please note, I have a BS and an MS degree in Electrical engineering
(in other words, 6 years of advanced engineering mathematics). My Ph.D. is
in mathematical Psychology (and when I switched to the Psych department
from EE, my advisor said that I didn't know enough math. (He, Duncan Luce,
had a Ph.D. in math). He made me learn topology.
In principle, I am a firm believer in the power of mathematical modeling.
My first psychology publications were math models, including the book
"models of human memory." I developed stochastic models of learning and
later on, math models of signal detection theory for application to the
study of human memory. The standard measure now frequently used in
healthcare of Area under the operating characteristic came from my early
work.
But the longer I work in this field, the more I am convinced it cannot be
done. Note that of the early large batch of mathematical psychologists,
most of them stopped doing math models. I like to say that we started with
a tool and looked for problems. Most of us realized the better way to
advance a field is to start with the issues, then figure out how to address
them. Math was seldom relevant.
Every year, I become less formal and more qualitative. But I still insist
on rigor: evidence-based. I do see some promise in agent-based computer
models, in dynamical systems (such as neural networks and related
approaches) and in other approaches.
But there are far too many variables that govern human behavior for us to
know how to model real behavior. Can it ever be done? Perhaps, but not in
my lifetime.
Terry also asked:
On Sun, Sep 10, 2017 at 6:03 AM, Terence Love <[log in to unmask]> wrote:
>
> Wondering how and when you see this developing in the context of your
> current program? Also would that then provide a basis for design to be
> included in National Science funding? And if so, when do you see it likely
> and what would be the readiness indicators?
>
In my carer, Ihave had many NSF grants. I have been on many NSF review
panels, including the advisory panel for CISE (Computer science) at NSF
that advises on principles and strategy.
I am in the process of submitting an extremely large NSF grant (>$20
milion), partnering with the contextual robotics group at UCSD (which is a
group housed jointly in engineering and social science and with several
Design Lab faculty members of the robotics groups as well as ours).
Evidence-based technology and science is completely acceptable within NSF.
They also have a large funding program in the behavioral sciences, where
much of the work is experimental: evience based, but not necessarily using
math or computer models. Experimental, however, means heavy and deep
reliance on evidence.
This is what we need more of in Design: Evidence-based principles.
don
Don
--
Don Norman
Prof. and Director, DesignLab, UC San Diego
[log in to unmask] designlab.ucsd.edu/ www.jnd.org <http://www.jnd.org/>
-----------------------------------------------------------------
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
-----------------------------------------------------------------
|