Brief and simultaneous assessment of delay, probability, and joint delay–probability discounting: Initial validation of the three-for-one discounting task in tobacco use [preprint]

<p>AbstractWhile delay discounting (DD) is a reliable marker of tobacco use, far less is known about probability discounting (PD) or how delay and risk jointly influence decision-making. To allow simultaneous assessment of DD, PD, and their interaction, we developed the three-for-one (TFO) discounting task. In 36 trials (&amp;lt;3 minutes), the TFO task estimates Effective Delay 50 (ED50) across a range of reward probabilities, allowing both empirical analysis of how risk impacts DD and quantitative modeling to derive independent DD and PD rates (k, h). Adults (n=134) reporting current- or never-cigarette smoking completed the TFO task, standalone DD and PD tasks, and a risk-numeracy scale (measuring understanding of risk/probability). The current vs. never-smoking group showed lower ED50 (higher DD) for certain/near-certain rewards (ps≤.021), but not riskier rewards. TFO-derived DD and PD rates showed evidence of known-groups validity (sensitivity to smoking-related DD differences, p=.002), discriminant validity (small correlation between DD and PD, r=.11), and convergent validity (moderate-to-strong correspondence with standalone tasks, ps&amp;lt;.001). The never-smoking group showed higher between-task correspondence in DD (p=.001); however, this effect was better accounted for by a moderating role of risk numeracy, evident in both DD and PD (ps&amp;lt;.001). Collectively, the TFO task showed that risk attenuates smoking-related group differences in DD and provided generally accurate assessment of DD and PD rates. Accuracy may be lower in participants with lower risk numeracy, but this does not impact sensitivity to known group differences in DD. Future work should explore additional validation approaches while evaluating methods to mitigate numeracy-related effects.</p>