Assessment & Research

Accurate characterization of delay discounting: a multiple model approach using approximate Bayesian model selection and a unified discounting measure.

Franck et al. (2015) · Journal of the experimental analysis of behavior 2015
★ The Verdict

Use free Bayesian code to pick the best delay-discounting curve per client instead of assuming one size fits all.

✓ Read this if BCBAs who assess impulsivity or self-control with delay-discounting tasks.
✗ Skip if Practitioners who only record response rate or latency, not choice across delays.

01Research in Context

01

What this study did

Alaimo et al. (2015) built a computer tool that picks the best delay-discounting curve for each person.

Instead of forcing every client into the same math model, the tool tests five common curves and chooses the one that fits best.

It also spits out a single number called ED50 so you can compare clients even when their curves differ.

02

What they found

The paper shows how to run the tool, not new client data.

It gives step-by-step code and worked examples so you can copy-paste into your own analysis.

03

How this fits with other research

Gilroy et al. (2018) took the same Bayesian picker and added a smart way to calculate area-under-the-curve when ED50 can’t be computed.

Franck et al. (2019) wrote a follow-up tutorial that walks you through priors and credible intervals, building on the 2015 tool.

Yeh et al. (2025) looked at a different problem—risk plus delay—but skipped the Bayesian picker, showing the field still hasn’t agreed on one gold-standard method.

04

Why it matters

If you run delay-discounting assessments for impulsive clients, stop guessing which curve to use. Let the Bayesian tool pick for each person, then report ED50. You’ll get cleaner data and easier comparisons across studies.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Download the T et al. script, feed in last week’s discounting data, and replace your old curve fit with the Bayesian-chosen one.

02At a glance

Intervention
not applicable
Design
methodology paper
Sample size
111
Population
neurotypical
Finding
not reported

03Original abstract

The study of delay discounting, or valuation of future rewards as a function of delay, has contributed to understanding the behavioral economics of addiction. Accurate characterization of discounting can be furthered by statistical model selection given that many functions have been proposed to measure future valuation of rewards. The present study provides a convenient Bayesian model selection algorithm that selects the most probable discounting model among a set of candidate models chosen by the researcher. The approach assigns the most probable model for each individual subject. Importantly, effective delay 50 (ED50) functions as a suitable unifying measure that is computable for and comparable between a number of popular functions, including both one- and two-parameter models. The combined model selection/ED50 approach is illustrated using empirical discounting data collected from a sample of 111 undergraduate students with models proposed by Laibson (1997); Mazur (1987); Myerson & Green (1995); Rachlin (2006); and Samuelson (1937). Computer simulation suggests that the proposed Bayesian model selection approach outperforms the single model approach when data truly arise from multiple models. When a single model underlies all participant data, the simulation suggests that the proposed approach fares no worse than the single model approach.

Journal of the experimental analysis of behavior, 2015 · doi:10.1002/jeab.2