38
M. Wilke et al.: Risk Specificity for Risk-Defusing Operators
ticipants requested information about positive and negative
consequences. This seems to be a result of the construction
of the scenarios and from the general understanding of risks.
Cost-benefit notions can be seen as an elementary part of
decision making, and this seems to be quite independent of
the type of risk with the exception of normal and global
risks, for which negative consequences were more impor-
tant.
Significantly more participants asked questions about
positive consequences for normal risks than for any other
type of risk. This over-representation of the normal type of
risk seems to be consistent with the possible outcomes of
these risks. In the normal area, decision makers want to
know about the positive chances that are inherent in the
risks. This “looking for a chance” might be easier if the risk
is not as high as other types of risks because of the poten-
tial reversibility of actions. That matches the significantly
smaller frequency of this type of statement for catastroph-
ic risks. For such risks, positive aspects are less apparent at
first glance. With respect to the alternative selected in the
end, a preference for the more secure alternative in cases of
increasing extent of damage and decreasing degree of re-
versibility was shown, corresponding with the classifica-
tion criteria.
Final Reflections
The combination of C-AIS, the thinking-aloud method, and
a post-decision interview was chosen to collect a broad
spectrum of verbal data. The comparison of the mean num-
ber of questions posed and statements made shows that near-
ly half of the verbal information collected came from think-
ing aloud. Therefore, this method can be seen as an
important source of information and as a useful addition to
theAIS. Especially the new category “attitudes/rules/plans”
seems to play an important role in the context of global risks.
Further research using this multi-method approach seems
worth the high investment of time, which is needed for tran-
scription.
The results of the present study do not reveal whether a
specific type of risk would allow the prediction of RDO
variance. Because risk domain and the specific combina-
tion of domain and type are influential as well, it seems
more promising to construct an even more detailed typolo-
gy including formal and content criteria that could explain
variations in decision-making behavior. This is not an ar-
gument against a formalization of risks, but it points to the
importance of a detailed analysis of risk situations. This is
in accordance with the naturalistic approach to decision
making in which concrete situations, partly single case stud-
ies, are the focus of the research.
Some researchers (e.g., Brase, Cosmides, & Tooby,
1998; Cosmides & Tooby, 1987; Gigerenzer & Goldstein,
1996) assume that humans’ cognitive machinery has had to
adapt to many different situations during evolution; there-
fore, no one general, but many problem-specific, adaptive
processing mechanisms have developed. It follows that on-
ly a detailed model, which integrates many problem- and
situation-specific aspects, can map real decision problems.
It remains open whether further factors beyond the ones
analyzed here (type of risk, risk domain, and the interaction
of the two) affect the preferences for certain RDOs. So the
scope of the RDO concept must be more clearly defined.
The intention of our study was to appraise non-experts’ as-
sessment of risks and to investigate how critical aspects of
risks can be reduced. For all types of risk, the decision mak-
ers tried to precisely explore the situation and the potential
negative consequences. These components seem to be an
important condition for evaluating and planning. Sufficient
information could mean a safety for the further handling of
risks. If an RDO is supposed to reduce negative conse-
quences, then these consequences have to be investigated
closely. For practical purposes, this means that the oppor-
tunity of the risk can be used by adequate strategies as far
as a priori calculations are possible. There is evidence that
the frequency with which the more risky alternative is se-
lected decreases from normal to medium to catastrophic to
global type of risks, whereas the amount of detailed infor-
mation about consequences decreases, too, due to increas-
ing complexity.
Well-founded education about risk-taking should be sup-
ported and put in place specifically for global risks with
long-term effects. There seems to be a high need for infor-
mation because non-experts fall back on existing and less
well-founded attitudes (e.g., “I have preferred organic prod-
ucts for years, so I would never support genetically altered
food”). Under the condition of sufficient information, ac-
tive risk-defusing could be supported accurately. The gen-
erally preferred strategy (within all four types of risk) of
new alternatives shows that it could make more sense to
look for new ways - even for risks that are extremely threat-
ening - than to restrict the damage afterwards or before (par-
ticipants often suggested the search for “something else,”
e.g., another kind of rubbish dump, another kind of stock).
In addition, detailed and comprehensive worst-case plans
should be available for catastrophes (e.g., participants in-
tensively explored information about the side effects of
medicine). Participants showed a good feeling for situations
in which preventive strategies could not guarantee risk re-
duction but where, instead, compensation strategies were
helpful. As for global risks (characterized by long-term ef-
fects), participants proposed that time be used effectively
to create controlling strategies until the triggering event oc-
curs (e.g., they suggested that genetically modified food
should be labeled). In this case, simulation studies referring
to possibilities for control would be very useful for dealing
with complexity and the long-term effects of such risks (par-
ticipants requested simulation studies very often in the in-
terviews). As an aim for upcoming research, such consid-
erations could be investigated in detail to support efficient
risk management.
Swiss J Psychol 67 (1), © 2008 by Verlag Hans Huber, Hogrefe AG, Bern