Additional model-free measures of discounting behavior are correlated with the discount factors estimated using BHM model. Each dot is one subject from the main experiment. (A, B) Total time spent per task (x-axis) plotted against fraction of total reward (y-axis) shows significant positive relationship (NV: Pearson , ; SV: Pearson , ). The more you wait (the more you spend time for the session), the more you earn, unless you make first-order violations by choosing later offers that are worse than the immediate option. (C, D) The logs of discount factors (x-axis) plotted against the fraction of total reward (y-axis) result in significant negative correlations for NV (Pearson , ) and SV (Pearson , ). (E, F) The logs of discount factors (x-axis) plotted against the non-waiting time (total task time minus waiting time, y-axis). In NV Pearson , and SV Pearson , . Similar significant relationship is also observed for the and waiting time (not plotted, NV: Pearson , ; SV: Pearson , ). What is important, across subjects non-waiting time is significantly correlated with waiting time, meaning that impulsive subjects were also quick in everything else (reaction time, mouse movement and clicks, NV: Pearson , ; SV: Pearson , ). (G, H, I) Significant negative relationship of the logs of discount factors (x-axis) with the total time (total time spent on a task, y-axis) in the NV (Pearson , ) and SV (Pearson , ), but not LV (Pearson , ).