The Effect of Numeric Format on Probability Discounting Rates of Medical and Monetary Outcomes

We assessed whether numeric format in the form of frequencies and percentages affects probability discounting rates in medical and monetary contexts across three experiments. Experiments 1 and 2 compared percentage (e.g., 10%) and frequency (e.g., 10 in 100) formats. Numeric format only affected discounting rates in the medical condition in Experiment 2. The lack of effect of numeric format on probability discounting of money was inconsistent with previous findings from Yi and Bickel (2005). Therefore, in Experiment 3 we more closely replicated their procedure. Results yielded no effect of numeric format on discounting rates. Across experiments, we concluded that probability discounting rates for monetary outcomes are not affected by whether they are presented in a frequency or percentage format. We also concluded that the effect of numeric format on probability discounting rates of medical outcomes is fragile, and unlikely to affect medical decision making in practice. Experiment 2 and 3 also examined whether the hyperbolic or hyperboloid model provided a more efficient description of probability discounting functions. Data were more efficiently described by the hyperbolic model in Experiment 2, and the hyperboloid model in Experiment 3.