|Explorations of the Magnitude Effect in Various Species and Domains
|Sunday, May 28, 2017
|8:00 AM–9:50 AM
|Hyatt Regency, Centennial Ballroom B/C
|Area: EAB; Domain: Basic Research
|Chair: Charles Frye (Utah State University)
|Discussant: Suzanne H. Mitchell (Oregon Health & Science University)
|Abstract: Discounting refers to the decline in value of a reward with increasing cost associated with the receipt of that reward. When the cost is “delay” humans discount relatively small rewards to a greater degree than relatively large rewards. We call this phenomenon the magnitude (or amount) effect. While the magnitude effect has been reliably observed in human delay discounting research using monetary gains, less attention has been paid to the magnitude effect in other discounting situations and other species. Recent investigations of the magnitude effect across various species and domains will be discussed in this symposium. Topics include: (1) The influence of context and framing on the magnitude effect in delay discounting using capuchin monkeys as subjects (Smith & Beran). (2) Differential effects of magnitude on gains and losses in delay and probability discounting in humans (Green & Myerson). (3) Investigation of the magnitude effect in delay, probability, and effort discounting with humans (Escobar & colleagues). (4) The effect of within-session contrast on the magnitude effect in delay discounting with humans and pigeons (Frye & Odum).
|Instruction Level: Basic
|Keyword(s): delay discounting, effort discounting, magnitude effect, probability discounting
|The Magnitude Effect in Nonhuman Primates
|TRAVIS RAY SMITH (Georgia State University), Michael J. Beran (Language Research Center, Georgia State University)
|Abstract: Two experiments assessed whether the magnitude effect occurs in nonhuman primates. In Experiment 1, capuchin monkeys chose between a smaller-sooner (SS) outcome and a larger-later (LL) outcome, where the LL outcome was always delayed 10 s and delivered twice the food pellets. Three session-types presented (A) 1v2, (B) 4v8, or (C) 2v4 pellets. The design involved two consecutive daily sessions where session type A or type B always started the block and session type C always followed the next day. This design should frame the session-C choices such that the delayed option is chosen more often following session-A (small reward) opportunities than session-B (large reward) opportunities (Dai, Grace, & Kemp, 2009). In Experiment 2, macaques chose between the SS reward and a variety of LL rewards (2, 4, & 8, pellets) across a variety of delays (2, 4, 8, & 16 seconds), varied across sessions. The SS reward amount titrated based upon preference in the prior 4-trial block, a design used to determine whether the discounting rate for delayed rewards decreases as the magnitude of the LL reward increases (Green, Myerson, Holt, et al., 2004). Thus far, neither experiment has shown conclusive evidence of a magnitude effect in monkeys.
Differential Effects of Outcome Magnitude on Discounting of Delayed and Probabilistic Monetary Gains and Losses
|LEONARD GREEN (Washington University), Joel Myerson (Washington University)
Much less is known about the discounting of delayed losses than about the discounting of delayed gains. We present our findings from recent studies involving the discounting of losses, including one conducted in our laboratory that examined discounting of both delayed losses and probabilistic losses over a wide range of amounts. In another conducted online, we examined discounting of delayed losses using a questionnaire modelled after Kirby's monetary choice questionnaire. The results from the laboratory study revealed no consistent effects of magnitude with either delayed losses or probabilistic losses, in contrast to the well-established finding of robust magnitude effects with delayed gains, in which larger amounts are discounted less steeply than smaller ones, and also with probabilistic gains, in which larger amounts are discounted more steeply than smaller ones. The results from our online studies replicate these differential magnitude effects for delayed gains and delayed losses using a different procedure and much larger, more representative samples. Taken together, these findings argue that although delay and probability discounting involve fundamentally different decision-making mechanisms, nevertheless the discounting of delayed and probabilistic losses share an insensitivity to amount that distinguishes it from the discounting of delayed and probabilistic gains.
Effort Discounting and its Relationship With Delay Discounting
|GISEL GÓMEZ ESCOBAR (University of Guadalajara), Cristiano Dos Santos (Universidad de Guadalajara), Amy Odum (Utah State University)
The evidence that delay and probability discounting may be different processes in decision-making is robust. However, the findings about effort discounting and its possible relationship with delay discounting are unclear, because the results seem to depend on the framing of choice, as well as the participants. The present study evaluated the magnitude effect in delay, probability, and effort discounting with hypothetical outcomes in psychology major college students . The results suggest that the magnitude effect in effort discounting is similar to the effect in delay discounting, namely, that larger rewards are discounted less steeply than smaller rewards. The hyperbolic model was also able to fit the data for the effort discounting task. We propose to apply Structural Equation Modeling for future studies to assess possible underlying processes in choice behavior that includes delays, probability, and effort required for outcomes. Furthermore, different procedures could give information about under which circumstances delays and effort discounting are similar or different from each other. Regardless, in a applied context, effort discounting would be useful to identify patients with potential low adherence to treatment.
|Examining the Magnitude Effect in Humans and Pigeons: It is All About the Contrast
|CHARLES FRYE (Utah State University), Amy Odum (Utah State University)
|Abstract: Delay discounting refers to the decline in value of a reward with increasing delays to receipt of that reward. Typically, humans discount relatively small rewards to a greater degree than relatively large rewards (i.e., the magnitude effect). Typically, non-humans do not discount rewards of different objective magnitudes to a different degree (i.e., non-humans do not show a magnitude effect). The difference between humans and non-humans in regards to illustration of the magnitude effect has been suggested to be a species difference. However, a procedural difference in human and non-human studies investigating the magnitude effect has not been fully explored and may be of theoretical importance. Specifically, most experiments examining the magnitude effect in human participants expose participants to both magnitudes in the same session, whereas most experiments examining the magnitude effect in non-human subjects hold magnitude constant within session, and manipulate magnitude once some stability criterion has been reached (i.e., the two magnitudes are not experienced in the same session). Across two experiments we examined the contribution of within-session contrast on the magnitude effect in humans (Experiment 1) and pigeons (Experiment 2). Both humans and pigeons showed a more robust and reliable magnitude effect when both magnitudes were experienced in the same session.