|
| On the Predictive Utility of Discounting Models |
| Sunday, May 25, 2025 |
| 9:00 AM–9:50 AM |
| Marriott Marquis, M2 Level, Marquis Salon 1-5 |
| Area: EAB/SCI; Domain: Basic Research |
| Chair: Darlene E. Crone-Todd (Salem State University) |
| CE Instructor: David J. Cox, Ph.D. |
| Presenting Author: DAVID J. COX (Endicott College; Mosaic Pediatric Therapy) |
| Abstract: The traditional process of conducting quantitative analytics in discounting research involves fitting one or more mathematical models of choice to the indifference points obtained for each participant, often using monetary outcomes. Here, the empirical goal is often to see how well the model fits the data and what the interpreted parameters might mean for the basic or theoretical question under study. In contrast, quantitative analytics in other scientific domains involves fitting one or more mathematical models to only a portion of obtained data to see how well the model can predict data unavailable during model building. This approach is particularly useful when the model is deployed in everyday situations where predictions about choice are needed across many unique decision contexts. In this presentation, we review a series of human operant experiments and computer simulations that robustly test the ability of discounting models to predict participant choices outside of the task from which the participant discounting rates were derived. This approach to testing the predictive utility of quantitative analytics in discounting research offers an alternative method whereby translational researchers can continue developing methods that make discounting models more useful for predicting human choice in socially significant situations. |
| Instruction Level: Intermediate |
| Target Audience: Attendees should have a basic understanding of what discounting is. The presentation will briefly review the basics of model fitting in discounting so not a pre-requisite per se. But, having some familiarity with how this process currently works will be useful. |
Learning Objectives: 1. Describe the traditional, descriptive methods for fitting discounting models.
2. Describe one alternative, predictive method for using discounting data.
3. Understand the benefits and limitations of each approach for the predictive utility of discounting models. |
| |
| DAVID J. COX (Endicott College; Mosaic Pediatric Therapy) |
 Dr. David J. Cox has formal educational training in psychology (B.S.), bioethics (M.S.), behavior analysis (Ph.D.), behavioral pharmacology (post-doc), and data science (post-doc). For the past 12 years, his research methods and tools have focused on quantitative analyses of choice, spanning basic research in human operant experiments to artificial intelligence research using big data. His research goals have aimed at using mathematical models of choice to describe, predict, and improve humans' daily decisions. This work has led to the publication of 60+ peer-reviewed manuscripts and seven books. |
|
| |