Information Technology Reference
In-Depth Information
(outgoing influences). Further possibilities for model enhancement include making
use of questions about outgoing influences to define the expertise and number of con-
nections of an agent.
Social Interaction and Network:
Agents in the consumat approach are more likely
to interact the more similar they are. Similarity is based on data, such as age, income,
critical opinions, and on physical proximity in the model (neighborhood effect). This
will result in a network describing the frequency of interactions over time between
different agents, however will not lead to an explicitly dynamic network as most of
the variables determining the chance of interaction are fixed.
3
A Perspective on the Use of the Model: Towards Adaptive
Policy
The previous section showed on a general level the model components and its para-
meterization. Due to the high number of model components and of questions in the
dataset complete formalization would be too lengthy. Moreover, formalization is yet
ongoing. The envisaged aim of the model is testing of adaptive policy. Theory pro-
poses that in the process of diffusion (in this case, of electric cars), different consum-
ers have different sensitivities for technical and social drivers in the system. Hence
consumers that are most prone to adopt first are more likely to respond to different
drivers than consumers that are expected to be the last adopters. This implies that the
first group should be stimulated in a different manner than a potential second group of
adopters. Failure or success in stimulating the first group will have second order ef-
fects on the next groups of consumers. For example, if a first group has negative ex-
periences with range, this may be communicated to other consumers. Even if at a later
stage the range is improved, the negative experiences of the first group will as yet be
stored in the memory of other groups, and hamper the diffusion of electric cars as
long as these consumers are still satisfied with fuel cars. Policy should thus be target-
ing a particular segment of consumers that is likely to respond positively on specified
stimuli. Only after a critical part of this segment has adopted the product, a next group
can be targeted. Approach to a next group may involve a different set of stimuli. For
example the positive experiences of the first adopter group can be used as an argu-
ment, but the anti-conformist argument clearly becomes less relevant as diffusion
progresses while the conformist argument grows in importance. Also it is of interest
to whether the speed and the extent of diffusion are stronger influenced by incremen-
tal improvements of vehicle characteristics such as range and charging speed, or by
less frequent “breakthroughs” which generate more attention to the product. Our aims
with a fully parameterized model are thus the testing of the effects of both technologi-
cal improvements and communication strategies directed at stimulating the diffusion
of electric cars. First results are expected to be available for presentation at the
conference.
