Comparison of distortions of probability information in three stochastic tasks: Visual, visuo-motor and decision making under risk
Year: 2010
Type: article
Abstract: In decision making under risk (DMR), subjects deviate from Expected Utility Theory (EUT) by distorting probability information (Allais, 1953). We compare performance in DMR, a visual task equivalent to DMR and a visuo-motor task equivalent to DMR. In the visual and visuo-motor task, subjects shot at rectangular targets on a screen with a stochastic shooter. During a training phase (600 trials) subjects could learn how the probabilities of hitting a target varied with width. Following training, subjects were presented with pairs of targets. The widths of the targets were adjusted so that the subject could win points Oi with probability pi (i = 1,2) by choosing to aim at zone i. In 4 conditions, we varied the probability p2 of incurring outcome O2 between 0.8, 0.6, 0.4 and 0.2 by varying the size of one of the targets. The larger target always had value 1, the smaller, 2. We measured subjects' preferences by using an adaptive staircase procedure to estimate the probability p1 that corresponded to the point of indifference for choosing between the targets. In the visual task the subject selected a target by key press. In the visuo-motor task, the subject selected a target by touching it. The subject's own motor variance was effectively added to the shooter's variance in the motor task, decreasing the chances of hitting the target. In the third task (DMR) we displayed numbers representing probabilities and assigned outcomes of the two lotteries. Three subjects completed the experiment. The prediction of EUT is that p1/p2 at PSE should be constant, independent of p2 (Allais, 1953). We found that subject's preferences violated EUT in the DMR task, but not in the visual and visuo-motor task. In the latter two conditions we could not distinguish subjects' performance from optimal.
Source: Journal of Vision
DOI: 10.1167/9.8.1102
Language: English
Subfield: General Decision Sciences
Field: Decision Sciences
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $1,850
OpenAlex ID: W2016966548