Variations on the balloon analogue risk task: A censored regression analysis

Abstract

In the present project, we reexamined the balloon analogue risk task (BART) by evaluating three variations on the task: one that does not require pumping, one that controls for trial duration, and another that withholds feedback on popping until the end of each trial. To accurately assess the censored data produced by the BART, performance was compared across these variations using Bayesian analysis with censored regression. The first experiment compared a task that required pumping to one that did not, and revealed that the tendency to respond earlier than is optimal does not reflect an avoidance of effort. The second experiment included a condition in which the duration of each trial was held constant by continuing to automatically inflate a balloon to its maximum size after a cash-in response; feedback on the pop time was withheld until the end of each trial. This condition revealed that the tendency to respond earlier is not driven by a desire to finish the task quickly by cashing in early, but the results also strongly suggested that the immediate experience of popping created a greater aversion to risk (although this condition difference was inconsequential by the end of the experiment). The article concludes by considering the implications of these results for cognitive neuroscience approaches to understanding performance on the BART.

Keywords: Balloon analogue risk task (BART); Bayesian; Censored regression; Risk taking.