Behavioral results
For the pure-risk trials (low-risk, high-risk, mixed-risk) the ANOVA revealed a main effect of risk (F (2,36) = 5.98, p = .007, eta² = .25, epsilon = .92) on the percentage of choices of the option with higher variance of the outcomes (higher risk). Follow-up Helmert contrasts showed the significantly most cautious decisions in the mixed-risk condition as expected compared to the other two conditions (p = .016; Mean of mixed (M) = .35, SD = .16 and Mhigh = .44 and Mlow = .41, see Figure 1b). In addition, the percentage of cautious decisions was not significantly different between the low-risk and high-risk condition. There were no further significant effects.
For ambiguous trials, the ANOVA revealed a main effect for risk (F(2,36) = 12.86, p < .001, eta² = .42, epsilon = .86). It is important to note here that the spread or variance of outcomes was the same as in block 1 and defined the factor risk. As for the pure-risk trials follow-up Helmert contrasts showed the significantly most cautious decisions in the mixed-risk condition (p = .001, M = .33, SD = .18). Again, the percentage of cautious decisions was not significantly different between high and low-risk conditions. All other main effects and interactions were not significant.
Taken together, for both, the risk and the ambiguity block, mixed-risk options led to the highest percentage of cautious decisions. A combined analysis of risk and ambiguity blocks revealed a significant main effect of risk (F(2,36) = 15.7, p < .001, eta² = .47, epsilon = .84) and an interaction of risk with ambiguity (F(2,36) = 3.03, p = .023, eta² = .14, epsilon = .99). Post-hoc Helmert-contrasts showed that the interaction was due to more cautious decisions under high ambiguity for the mixed-risk condition compared to high-risk and low-risk trials in the conditions of pure risk and low ambiguity (p=.008). The condition with mixed-risk and high ambiguity led to the most cautious decisions (M = .30, SD = .18). Accordingly, high ambiguity further increased the risk-minimizing behavior in the context of mixed risk. Further, participants were relatively less cautious in the high-risk condition as compared to the low-risk condition under low ambiguity as compared to pure risk (p=.036). No other effects were significant.