Association for Behavior Analysis International

Reinforcers tend to systematically lose their value as the delay to their receipt increases - a process often called delay discounting. Because this devaluation process is particularly pronounced in clinical populations (e.g., substance abusers, obese individuals, problem gamblers) and the rapidity of delay discounting can predict treatment outcomes, considerable attention has been devoted to decreasing discounting rates. Although approaches such as framing (e.g., providing explicit dates or explicitly stating that choice of a reinforcer at one point precludes receiving it at another point) and delay training (i.e., providing histories of responding for delayed reinforcers) show promise the generality of these effects is unknown. The present symposium examined two challenges to the generality of these findings: their systematic replication, and the exclusion of non-systematic data.

People who steeply discount the future suffer higher rates of drug dependence, obesity, debt, and psychiatric disorder. Because steep temporal discounting precedes and predicts onset maladies like drug abuse, assessing discounting in clinical settings could improve prevention (e.g., identifying at-risk youth) and treatment (e.g., predicting relapse risk). Before crossing from laboratory to clinic, however, clinicians must know if temporal discounting can be faked by test takers motivated to respond deceptively (e.g., to avoid treatment). We measured, for the first time, the susceptibility of discounting assessments to deception. We adopted a common approach for measuring deception in psychological assessment, an analogue simulation design in which participants are given (a) standard instructions, (b) instructions to fake good (i.e., simulate better versions of themselves), or (c) instructions to fake bad. Significant findings revealed weaknesses of typical discounting assessments that should be addressed if these assessments are to have practical utility. Nonsignificant findings tentatively supported the use of discounting assessments in their current form.

Delay discounting (DD) and Probability Discounting (PD) are behavioral measures of impulsive choice. Discounting patterns tend to be predictable, where immediate (or certain) rewards are valued more than relatively delayed (or uncertain) rewards in a systematic way. However, some participants exhibit nonsystematic response patterns (NSR) that diverge from general expectations, which could have implications for the validity of discounting-related experiments. We meta-analyzed the findings from 114 discounting studies (from 78 total papers) that reported using Johnson and Bickels (2008) algorithms for identifying NSR and examined (1) the frequency of NSR patterns across studies; and (2) potential methodological factors that contribute to increased or decreased frequencies of nonsystematic responding. The overall frequency of NSR across DD and PD studies was 18% and 19%, respectively. Non-monetary outcomes (e.g., drugs, food, sex) yielded more NSR patterns (21%) than did discounting for monetary outcomes (16%; Q(1) = 3.87, p = .049). Our review indicates also that researchers are inconsistent in whether or how they report NSR in discounting studies, which is relevant for a clearer understanding of the behavioral mechanisms that underlie impulsive choice. We make several recommendations regarding the assessment of NSR in discounting research.

Temporal discounting is a behavioral process wherein positive outcomes are devalued as a function of delay. Recent research suggests that the framing of the delay (i.e., in days or dates) can modulate the degree to which human subjects discount the delayed outcome. Another framing effect involves the manner in which binary intertemporal choices are presented as discrete decisions or part of an extended sequence; specifically, explicitly stating that choosing a smaller-sooner reward is mutually exclusive to a later reward delivery (revealing a hidden zero), or implicitly implying so (keeping the zero hidden). The present study used an online-crowdsourced subject pool (mTurk; n=190) to assess delay discounting using a standard discounting task (Monetary Choice Questionnaire). Discounting was later assessed using the same task in a 2 (delay framing) by 2 (zero framing) design. Multiple regression suggested a main effect of delay framing after accounting for other demographic variables and manipulations, as well as a significant contribution of initial rate of discounting based on the standard Monetary Choice Questionnaire.