Information Technology Reference
In-Depth Information
-
Use the developed EDTM as the behavioural reference for the ABSS im-
plementation. This is pertinent as the EDTM is originally suggested to be
tested by administering the decision tree with the decision-makers. Thus
this adaptation proposes testing of the EDTM also at the ABSS level. The
advantage of this is to leverage the qualitative validation at the micro-level,
whilst contributing to cross-validate results at the macro-level. This valida-
tion approach is similar with the proposal discussed in [Moss and Edmonds,
2005].
-
As normality may not be an important issue in large datasets for forecasting
purposes [Maindonald and Braun, 2010], the testing of the developed EDTM
as an ABSS implementation is proposed to remain within the bounds of
whether the model results are deemed plausible by the stakeholders -without
necessarily providing predictions. Yet the model may still serve such purpose.
Figure 3 below depicts how the aforementioned adaptations interrelate in prac-
tice with the the EDAM development, discussed in the previous section.
administer follow-up
surveys and interviews
create a decision-making
model with behaviour rules
EDAM
steps
3 and 4
EDAM
steps
1 and 2
collect and analyse
ethnographic data
implement the resulting
EDTM as an ABSS
administer an intial survey
to the decision-makers
test both the
EDTM and the ABSS
EDAM
step 5
Fig. 3.
The adapted EDTM development cycle, following the EDAM
Once the surveys have been administered, one shall consider the ethnographic
approach -which can be done via semi-structured interviews and field observa-
tion. In terms of the ethnographic database, perhaps the most important aspect
is to make notes about the data as soon as it has been collected, preferably on
the spot, otherwise analyses can be adversely impacted by the increasing likeli-
hood of the qualitative data being misinterpreted, misremembered or forgotten
[Murchison, 2010]. Such detailed qualitative notes can be helpful in constructing
a narrative of what has been researched and also does help with fine tuning the
EDTM. One cannot record absolutely every detail, yet the modeller should have
enough information as to be able a clear understanding of the decision-making
process represented in an EDTM. As this is done from an ethnographic perspec-
tive, it may include -whenever appropriate and consented- recorded audio and
video as part of the database which grounds every step of the EDTM.
Due to the qualitative nature of the process, irrespective of whether the origi-
nal methodology or the adapted one is chosen for data collection, the usage of an
EDTM require modellers to take the following aspects into careful consideration:
