Representing Human Decision-Making in Environmental Modelling - Environmental Modelling: Finding Simplicity in Complexity - page 292

way to develop 'realistically descriptive specifications of
individual behaviour and social interaction', promote
learning and understanding, and aid negotiation pro-
cesses (Moss, 2008: abstract), it comes with high resource
requirements and demands unique skills on behalf of
the modeller.
As outlined above, agent-based approaches to mod-
elling human decision making seem to offer many advan-
tages over neoclassical economic approaches. However,
relative to the long history of economic modelling, agent-
based modelling is a field in its infancy that is only
beginning to establish theory and methods. A lack of
standard methods for model construction and analysis
or model comparison may have led some to question
the robustness of the approach for scientific analysis
(Windrum et al ., 2007). However, standard protocols
for describing (Grimm et al ., 2006, 2010), replicating
(Wilensky and Rand, 2007) and comparing (Polhill et al .,
2008) models are available, and the development of mod-
ular models may also aid development (Anon, 2009).
Other questions remain about agent-based model valida-
tion and assessment, whether done by comparing model
output with empirical data (e.g. Windrum et al ., 2007)
or with the direct involvement of the actors represented
(e.g. Moss, 2008; Millington et al ., 2011).
Although there may be a great allure to creating 'would-
be-worlds' to experiment on (Casti, 1996), agent-based
models will not be the best approach in all cases and as with
any modelling activity careful thought should be given to
the level of abstraction used to represent human decision-
making (O'Sullivan et al ., 2012). Agent-based models
can quickly become as complicated (and intractable) as
the system they intend to simplify (O'Sullivan, 2004),
requiring the modeller to think carefully about their
analysis and the purpose of the entire enterprise. However,
O'Sullivan et al . (2012: 119) suggest that ''Where agents'
preferences and (spatial) situations differ widely, and
where agents' decisions substantially alter the decision-
making contexts for other agents, there is likely to be a
good case for exploring the usefulness of an agent-based
approach.''

reason for this unpredictability is because socio-economic
systems are 'open' and have a propensity to structural
changes in the very relationships that we hope to model.
By open, we mean that the systems have flows of mass,
energy, information and values into and out of them
that may cause changes in political, economic, social and
cultural meanings, processes and states. As a result, the
behaviour and relationships of components are open to
modification by events and phenomena from outside the
system of study. This modification can even apply to us
as modellers because of what economist George Soros
has termed the 'human uncertainty principle' (Soros,
2003). Soros draws parallels between his principle and the
Heisenberg uncertainty principle in quantum mechan-
ics. However, a more appropriate way to think about
this problem might be by considering the distinction
Ian Hacking makes between the classification of 'indif-
ferent' and 'interactive' kinds (Hacking, 1999; also see
Hoggart et al ., 2002). Indifferent kinds - such as trees,
rocks, or fish - are not aware that they are being classified
by an observer. In contrast humans are 'interactive kinds'
because they are aware and can respond to how they are
being classified (including how modellers classify differ-
ent kinds of agent behaviour in their models). Whereas
indifferent kinds do not modify their behaviour because
of their classification, an interactive kind might. This
situation has the potential to invalidate a model of inter-
active kinds before it has even been used. For example,
even if a modeller has correctly classified risk-takers ver-
sus risk-avoiders initially, a person in the system being
modelled may modify their behaviour (for example, their
evaluation of certain risks) on seeing the results of that
behaviour in the model. Although the initial structure of
the model was appropriate, the model may potentially
later lead to its own invalidity!
This situation requires modellers to think carefully
about why they want to represent human behaviour in
their models. If prediction of a future system state is
the goal, a modeller might be best served by focusing
modelling efforts on the natural system and then
using that model with scenarios of human behaviour
to examine the plausible range of states the natural
system might take. Although some may see prediction
as the ultimate goal of any modelling (whether that is
prediction for hypothesis testing, forecasting the future
or otherwise), there are many other reasons to model (for
example, van der Leeuw 2004; Epstein, 2008). A more
heuristic approach might explore what types of human
behaviour or socio-economic structure are necessary or
contingent for alternative future system states (e.g. Sayer,

18.5 Discussion

At the outset of this chapter we highlighted the inher-
ent unpredictability of human behaviour and several of
the examples we have presented may have done little
to persuade you that current models of decision-making
can make accurate forecasts about the future. A major