Julian Newman (jne@gcal.ac.uk)
Wed, 20 Oct 1999 18:22:57 +0100
Date: Wed, 20 Oct 1999 18:22:57 +0100 From: Julian Newman <jne@gcal.ac.uk> Subject: Suitable Research Methods for Information Systems
List address to send message to everyone: ifets-discuss@LISTSERV.READADP.COM
Details of current discussion: http://ifets.ieee.org/discussions/discuss.html
---------------------------------------------------------------------------
On 14 Oct Sandra Windeatt <sandra.windeatt@unn.ac.uk> wrote:
> On 5th October, Julian Newman talked about the problems of applying
> Hypothetico-Deductive methods to systems development research
> projects.
>
> Is it worth investigating other paradigms such as Action
> Research? There is a lot of 'Soft Systems' literature on this
> method of solving information problems and developing information
> systems.
I have never really been happy about "Action Research", possibly
because of the personalities advocating it, or possibly because I felt
that the logic of 'Action Research' as research was confused or
unclear. (See
excerpt from Ania Lian below).
Sandra talks of Action Research/SSM as a "method of solving information
problems and developing information systems". My basic problem, though,
is not how to develop information systems but how to develop RESEARCH IN
information systems. Whereas in some subjects students learn how to
research (on a small scale) from the very beginning of their degree in
their weekly lab classes, in Computing the lab classes are oriented to
learning how to solve problems which are not research problems, they are
problems of design and implementation. So there is a big mismatch
between the skills they acquire in their practical classes and the
skills and way of thinking which will be needed if they are to do a
"research" project.
On 15 Oct 1999 Ania Lian <ania@lingua.arts.uq.edu.au> wrote:
> It seems to me that Action Research refers to a paradigm which assumes
> that even if you cannot find answers at some general level (in terms of a
> scientific formula), you can find formula (??) which work at some local
> level. The qualitative studies think that they do just that. The problem
> is that even qualitative studies, by virtue of methodology which imposes
> own values (bias) on selection and organisation of data are nothing else
> but quantitative. They just count things differently but still work on and
> present statistical data. In this sense, I see no reason for a division
> between any types of research.
I feel that up to a point this expresses my feeling, too. However, I do
see a problem in presenting and structuring all "Research Methods" as if
it were
aiming at the growth of scientific theory per se. If the practical
element of a research project is oriented to designing and/or
implementing a solution to a real-world problem, then it is almost
certainly not well designed as an experiment to test a hypothesis; and
since most of what is available to students by way of guidance presumes
that the growth of theory via hypothesis-testing is what we are about,
they just get confused. Not much chance of developing a rigorous test
of a hypothesis if you do not have a body of theoretical argument that
generates testable theoretical issues, and if you do not have the
experience of working in various experimental or quasi-experimental
paradigms.
Ania goes on to say:
> What I do see is a reason for
> examinaton of the research methods in terms of the goals and the
> principles selected for their realisation. In short, how is our research
> going to help our goals?
This is spot-on, and through time I have come to recognise the value of
distinguishing, as John Long has done, between "Science" "Engineering"
and
"Systems Development" paradigms. This is a useful way of thinking about
what the relationship may be between your research activity and the
"Real
World". The Hypothetico-Deductive approach gives a good guide to how to
do Science; but the result of doing that is basic, generalisable
knowledge, and not necessarily the solution to a practical problem. It
is universalistic
by dint of abstracting a great deal from messy, particular
circumstances. "Engineering" is also fairly universalistic, in that it
uses Science to produce generic solutions (Long says "functions") - e.g.
the lever, the jet engine etc - so it shows how we can typically act on
the world. "Systems
Development" takes the world as the source of Requirements - so gets
closer still to the messiness of particular circumstances. Clearly a
major hurdle that students have to overcome in beginning research is to
focus on clear, coherent and attainable goals. I think this is made
much harder because there are so few examples of IS researchers actually
having clearly-expressed goals in their published research!
Sandra also wrote:
> The "Hypothetico-Deductive philosophy of science" is not always
> "great for Social Science" either and there is also a huge literature
> about other (qualitative) methods, for example "evaluative methods"
> which may better "fit to the situation where the practical part of the
> study involves developing an artefact."
I know that many Social Scientists regard the Hypothetico-Deductive
method as inappropriate, too; however I tend to feel this is often the
result of not having given it a serious try (though I have to confess to
having published more qualitative/participant observation work than
quantitative work myself!) I agree that systematic evaluation is
probably the appropriate alternative to theory-testing where
implementation-oriented projects are concerned. But there does remain a
considerable problem of ensuring that the results are in some way
generalisable, else I don't think they can sustain a claim to be a
contribution to knowledge.
---------------------------------------------------------
Forum website: http://ifets.ieee.org/
Forum's contact person: kinshuk@massey.ac.nz
Info on Join/Leave List: http://ifets.ieee.org/maillist.html
---------------------------------------------------------
This archive was generated by hypermail 2.0b3 on Wed 27 Oct 1999 - 22:59:27 MEST