Abstract Delay discounting entails a choice between one outcome soon and a more valued outcome later. Analyses emphasizing delay mask the role of the outcomes’ magnitudes in shaping the discounting function. Discounting was examined for justice-involved youth (JIY) and college students. Individuals performed a delay-discounting task using a range of smaller-sooner (SS) magnitudes and larger-later (LL) delays. First, a discounting function was formed by mimicking an adjusting amount procedure. A hyperboloid equation modelled the data well, based on information-theoretic model-selection, but many datasets from the JIY had to be eliminated. Second, dichotomous choices (“wait,” “don’t wait”) were modelled using mixed-effects logistic regression that incorporated both SS magnitude and LL delay. All data sets could be included in the second approach, which was based on the concatenated matching law. Datasets that were eliminated from the hyperboloid analysis showed sensitivity to reinforcer magnitudes undetected by the hyperboloid model. JIY discounted at a higher rate than college students in both analyses, partly because of poor sensitivity to the outcomes’ relative magnitudes. The two groups were tested with different modalities, limiting the conclusions drawn about them. Nonetheless, for both groups and testing modalities, incorporating magnitude and delay in mixed-model analyses enhanced the generality of delay-discounting analyses.
Newland et al. (Tue,) studied this question.