Page 268 -

12 Participatory Approaches 267

Laboratory settings are very controlled experiments, involving human subjects.
This is the case for most economic experiments. Participants are encouraged to
behave with a given rationality through instructions and payments at the end of
the session. In canonical experiments, analysis of the experiments is performed by
the scientist. The focus of the analysis is to understand the individual and collective
behavioural patterns generated by these settings. The purpose of these experiments
is either (i) to test theories and models, (ii) to gain new knowledge on human
behavioural patterns in given situations, or (iii) to test new institutional configu-
rations (Friedman and Sunder 1994). These experiments are particularly efficient
for situations with strong communication issues or with important interindividual
interactivity (Ostrom et al. 1994). The issue of simulating a real situation is not
considered but rather the testing of a theoretical model. This field is currently very
active and evolves with the emergence of field experiments involving stakeholders
concerned by the issues idealised in the model tested, asking them to play in their
environment (Cardenas et al. 2000). With this configuration, interactions are rather
deep since participants act as parts of the model. The participants convey action
choices. However, the experimentalist strongly controls the process.
A platform is an intermediary setting more open to compromise and hybridisation
than the laboratory. Heterogeneity of participants is also more welcome, since
the setting is designed to enhance sharing interests. Through experimentation, a
platform is supposed to bridge the gap between the world of the model and that
of the stakeholders (Callon and Muniesa 2006). Policy exercises and role-playing
games, as developed in the companion modelling approach, are kinds of these
platforms (Richard and Barreteau 2006). Policy exercises embed stakeholders in
potential situations they might have to face in the future (Toth 1988). They stem
from war games that have been developed since the time of ancient China and are
now used in public policy assessment (Duke and Geurts 2004) or environmental
foresighting (Mermet 1993). They are actually quite similar to the business games
and the system dynamics trend explained previously in Subsect. 12.2.1. However,
the underlying social simulation model is rather implicit, though it exists to create
the potential situation and to help identify the participants relevant to the exercise.
Association with a computer tool tends to be with a simulation model of the
environment that does not necessarily involve a social component. The interaction
between participants and the social model is rather deep since they are pieces of
the model and connect with the model of their environment. Control of the process
is rather diffuse. There might be a genuine empowerment of participants since they
have the possibility of bringing their own parts of the social model to the process and
can adapt it in ways different to what the designers expected. Alike with laboratory
settings, platforms provide information to the modeller about behavioural patterns
of the participants. Reaction to taboos or innovative behaviours in situations new
to the participants, tacit routines and collective behavioural patterns can be elicited
using these platforms, while it is difficult with classical interviewing techniques.
Between experimental laboratory settings and policy exercises, the companion
modelling approach proposes an association of role-playing games and agent-based
simulations (Bousquet et al. 2002). Even though authors in this approach claim not