The study was designed to investigate the relative input of payoffs and probabilities into risk judgment on the basis of the analysis of information search pattern. The modified version of MouselabWeb software (http://www.mouselabweb.org) was used as an investigative tool. The amount, the kind and the order of information accessed by subjects to evaluate risk was collected from ordinary respondents and respondents trained in mathematics and statistics. In the latter group were 75 students and young researchers working at National Aeronautics and Space Administration (NASA). The ordinary subjects were 67 Polish students of social sciences. As expected, the NASA group considered more information than ordinary students and searched for more information about probabilities. However, the ratio of information about payoffs to probabilities was close to 1 in both groups. Moreover, average risk rates were similar in both groups. It was also observed that risk rates were positively related with the amount of information about probabilities considered by subjects.

The most salient characteristic of a risky choice is an approach-avoidance conflict-a desire to maximize benefits and minimize risk. This conflict is directly addressed in Risk-Value (R-V) models. Our goal is to determine to what extend the positive correlation between risk and value assumed in these models actually holds, as it is a necessary condition for tradeoffs. If people do not perceive a R-V relation as positive, there is no conflict, and decisions may be made on the basis of either criterion according to a lexicographic strategy, whereby either the least risky or the most beneficial option is chosen. The dependence between perception of the R-V relation and decision strategies was tested in three studies performed with 411 university students. Experimental (Studies 1-3) and questionnaire (Studies 2-3) measurements were used. In experiments, participants rated riskiness of two-outcome lotteries and made choices between a lottery and a sure thing. The perceived R-V relation was also investigated by way of three scales in the Domain-Specific Risk-Taking (DOSPERT) questionnaire-willingness to take risk (RT), perceived risk (PR), and expected benefits (EB). Our findings support the hypothesis that perception of a positive R-V relation is necessary for trade-offs-the PR and EB scores are positively correlated for those, who make trade-offs. When the scores on the PR and EB scales are negatively correlated, preferences are described as a function of either perceived risk, or EB only.

No abstract