Uncovering individual learning about risk using preferred outcome distributions

The way how humans act in an uncertain situation, especially when it involves potential monetary gains or losses, has been agitating minds of many economists, psychologists and mathematicians for the last half a century. Recent research suggests that by using a graphical representation it is possible to communicate about investment risks more efficiently. Bas Donkers, Carlos Lourenço, Daniel Goldstein and Benedict Dellaert have conducted a Distribution Strategy Game (Donkers et al, 2013) that uses the intuitive interface of the Distribution Builder to assess players’ financial decisions under risk. The participants use the interface to construct their preferred outcome distributions, while being exposed to a limited budget. The current paper builds on the data gathered from 144 respondents that took part in the game. The analysis has been conducted to estimate the changes of players’ strategy throughout the game, as well as to assess their behavior depending on the risk profile. The results show that the strategy of the players is affected by the state-of-the-world scenario, which is reflected in more cautious tactics in case of low returns. There have also been signs of strategy adjustments as the game progressed in order to maximize the financial outcomes. The participants were split into two distinctive profiles – risk-seeking and risk-averse – based on their relative performance, 83% of which remained in the same group for the whole game. It has been pointed out that the more unfavorable conditions are, the more risk-averse the players become.