36
M. Wilke et al.: Risk Specificity for Risk-Defusing Operators
risks, a significant positive λ coefficient was found for new
alternatives: λ = 1.39, p < 0.05. In accordance with our ex-
pectation, more participants formulated questions and state-
ments corresponding to new alternatives than for the other
two RDOs.
For catastrophes, a significant positive λ coefficient of
1.19 (p < 0.05) was found for new alternatives. Contrary to
our expectation, the number of persons with questions or
statements corresponding to this RDO was higher than for
worst-case plans (λ = 1.01, ns).
For global risks, a significant λ coefficient of 1.61 (p <
0.05) was found for possibilities for control. In line with ex-
pectations, most participants produced at least one question
or statement concerning this RDO.
λ coefficients of the different RDOs are shown in Fig-
ure 1, separated according to the four types of risks. Even
if the results are not consistent with the hypotheses in every
respect, there are still significant differences in the interest
in various kinds of RDOs used for different risks. There-
fore, we accept the general hypothesis of risk specificity of
RDOs.
In addition, we investigated for which type of risk a giv-
en type of RDO received the most questions: New alterna-
tives was the most frequent for medium risk (λ = 1.39; p <
0.05), possibilities for control most frequent for global risk
(λ = 1.61; p < 0.05), and worst-case plans for catastrophic
risk (λ= 1.01; p< 0.05). Moreover, the interactions between
risk and contents in relation to RDOs are interesting (as the
following three parameters were redundant in the log lin-
ear model, no probabilities will be given). New alternatives
(λ= 6.24) and possibilities for control (λ= 7.18) were most
often produced under the combination of global risk in the
ecology domain. Worst-case plans were most often men-
tioned in the ecology domain (λ = 3.69).
-1,5
1,5
0,5
0
-0,5
1: normal 2: medium 3: catastrophes 4: global
Type of Risk
Figure 1. λ coefficients for the three types of risk-defusing ope-
rators new alternatives, possibilities for control, and worst-case
plans; separated for the four types of risk (normal, medium, cata-
strophic, and global).
In total, there is clear evidence for risk specificity of ac-
tive risk-defusing behavior. The combination of type of risk
with risk domain has the greatest influence on search for
RDOs.
Explorative Data Analysis
We analyzed the difference between the number of partic-
ipants with at least one question or statement in the cate-
gory background knowledge/specific previous knowledge
via χ2 tests. Frequency differences between types of risks
were not significant for this category, χ2 (3, N= 79) = 1.86.
In the category attitudes/rules/principles, a χ2 (3, N= 55) =
15.04, p ≤ 0.05, was significant. The standardized residu-
als for normal types of risk (-2.37) and global types of risk
(3.04) showed that this category was significantly infre-
quent and frequent, respectively, which makes sense: The
more risky and therefore dangerous the decision, the more
attitudes affect decision making. There were no differences
for the category situation, χ2 (3, N = 321) = 1.21. But the
questions on probabilities significantly deviated from
chance, with the χ2 (3, N = 166) = 15.64, p ≤ 0.05, and a
standardized residual of -3.34 for the global type of risk.
Negative consequences were homogenously distributed
over all four types of risk (χ2 (3, N = 323) = 0.43, ns). Pos-
itive consequences showed significant deviations with a χ2
(3, N= 201) = 8.61, p ≤ 0.05, and standardized residuals of
-2.2 with Type 3 catastrophic risk and 2.01 with Type 1 nor-
mal risk. For Type 3, there were fewer statements on posi-
tive consequences, in Type 1 more statements. Table 3 sum-
marizes the result of these exploratory research questions.
Participants’ tendency to choose the more risky option
decreased as type of risk increased. For normal risks, 71%
chose the more risky option as compared with 58% for medi-
um risks, 56% for catastrophic risks, and only 37% for glob-
al risks (χ2 (3, N= 222) = 10.61, p ≤ 0.05).
Discussion
Risk Specificity of Active Risk-Defusing
Behavior
The main question of our study was related to the existence
of RDO and their connection to different types of risk. We
will start the discussion with our hypotheses concerning
these relationships. For the normal type of risk, good un-
derstanding and management of these risks did not lead to
Swiss J Psychol 67 (1), © 2008 by Verlag Hans Huber, Hogrefe AG, Bern