M. Wilke et al.: Risk Specificity for Risk-Defusing Operators
31
sider relevant when rating risks. In our pilot study, we asked
20 participants (eight female; mean age 26 years) to rate 57
risk situations using the WBGU criteria. Each risk situation
was to be rated along the six dimensions mentioned above
on a scale from 0 to 4 (= maximum). Using cluster analy-
sis, we found evidence for four different types of risk, which
can be described as follows (whereby the order roughly fol-
lows an intuitive classification on the basis of event sever-
ity):
1. Normal risks: calculable risks involving damage that is
expected to be reversible and whose extent is low (e.g.,
German measles, light flooding, borrowing money).
2. Medium risks: risks with a medium rating based on the
WBGU criteria (e.g., carcinogenic substances like as-
bestos, waste disposal, stock investments).
3. Catastrophic risks: risks involving a great deal of dam-
age for which the time lag between the triggering event
and its consequences is short (e.g., avian flu pandemic,
volcanic explosion, terrorist attack).
4. Global risks: risks involving a great deal of damage that
are expected to be irreversible and have a very long re-
moval period (e.g., genetically modified food products,
extinction of endangered species, global warming).
This four-cluster solution portrays the classification of risks
according to the understanding of non-experts. Besides
these four types of risks, three different domains were dis-
tinguished: (a) ecology (e.g., flooding), (b) health (e.g., car-
cinogenic substances), and (c) economics/politics (e.g., ter-
rorist attack). In the following, we will analyze the
relationship between the three risk domains and the four
types of risk.
Main Study
On the Risk Specificity of Active Risk-
Defusing Behavior
The present paper discusses the question of the risk speci-
ficity of decision-making behavior in quasi-realistic sce-
narios. Decision making is similar to problem solving in
that the construction of a mental representation of the giv-
en situation is essential. The mental representation resem-
bles a complex system with causal relations between ex-
ternal events and consequences. Relations involving the
decision maker’s activities imply a kind of controllability
and are of particular interest. They negate the causal con-
nection between the risk situation and its negative conse-
quences. We need to investigate under what conditions what
kind of relationship (i.e., RDO) is preferred (Huber, 2004).
The central questions are: (a) Does the search for an RDO
depend on the type of risk situation? and (b) What type of
RDO will be searched for? More precisely, do decision mak-
ers search for different RDOs in different types of risk sit-
uations?
In our attempt to answer these questions, we relied on
Huber et al.’s (2001) threefold differentiation of RDOs: new
alternatives, possibilities for control, and worst-case plans.
By analyzing the relationship between the RDO search
process and clearly defined types of risk, we will specify
and extend Huber’s approach: The scope and relevance of
the theory, which up to now has only been tested in normal
risk situations, will be evaluated for more global risk situ-
ations.
Main Research Questions and Hypotheses
Based on the typology of risks presented above, we formu-
lated the following general hypothesis: The interest in dif-
ferent RDOs during decision making depends on the type
of risk specified. In other words, while participants think
aloud, the type and number of questions they pose and state-
ments they make about different RDOs (new alternatives,
possibilities for control, or worst-case plans) will vary de-
pending on type of risk.
As for the use of RDOs, we have specific expectations
for the different types of risks. The combination of the clas-
sification criteria (see pilot study) resulted in the following
hypotheses for each type of risk:
1. We assumed that decision making would be easiest for
the normal type of risk because it is calculable, involves
damage that is reversible and whose extent is low. This
type of risk was viewed as a kind of control group. For
this type of risk, we did not expect to observe specific
preferences for one of the RDOs.
2. In the event of a medium type of risk, in other words,
when decision makers expected the damage to be of
medium extent, largely restorable, to have a medium-
term removal period, a medium-level certainty of as-
sessment, and the time lag between the triggering event
and the damage to be medium-long, we predicted that
decision makers would request alternatives to the options
provided. This search for alternative options is very sim-
ilar to the RDO new alternatives. We therefore expected
questions and statements to most frequently concern new
alternatives.
3. We assumed that there would be a high amount of pres-
sure to act in light of a catastrophic type of risk, in oth-
er words, when decision makers expected the damage to
be extensive and only partially restorable, and the time
lag between the triggering event and the damage to be
short. In this case, decision makers may judge the po-
tential for reducing or eliminating negative conse-
quences to be very low. We thus predicted that they would
focus more on compensating for potential negative con-
sequences. Accordingly, we expected questions and
statements to most frequently involve worst-case plans.
4. For the global type of risk, in other words, when deci-
sion makers expected the damage to have a very long re-
moval period, to be irreversible, and the situation high-
ly complex, we predicted that decision makers would be
interested in coping with negative consequences by
Swiss J Psychol 67 (1), © 2008 by Verlag Hans Huber, Hogrefe AG, Bern