R2 should not be used to describe behavioral-economic discounting and demand models.

Literature concerning operant behavioral economics shows a strong preference for the coefficient of determination (R2) metric to (a) describe how well an applied model accounts for variance and (b) depict the quality of collected data. Yet R2 is incompatible with nonlinear modeling. In this report, we provide an updated discussion of the concerns with R2. We first review recent articles that have been published in the Journal of the Experimental Analysis of Behavior that employ nonlinear models, noting recent trends in goodness-of-fit reporting, including the continued reliance on R2. We then examine the tendency for these metrics to bias against linear-like patterns via a positive correlation between goodness of fit and the primary outputs of behavioral-economic modeling. Mathematically, R2 is systematically more stringent for lower values for discounting parameters (e.g., k) in discounting studies and lower values for the elasticity parameter (α) in demand analysis. The study results suggest there may be heterogeneity in how this bias emerges in data sets of varied composition and origin. There are limitations when using any goodness-of-fit measure to assess the systematic nature of data in behavioral-economic studies, and to address those we recommend the use of algorithms that test fundamental expectations of the data.

A systematic review and meta-analysis of test-retest reliability and stability of delay and probability discounting.

In this meta-analysis, we describe a benchmark value of delay and probability discounting reliability and stability that might be used to (a) evaluate the meaningfulness of clinically achieved changes in discounting and (b) support the role of discounting as a valid and enduring measure of intertemporal choice. We examined test-retest reliability, stability effect sizes (dz; Cohen, 1992), and relevant moderators across 30 publications comprising 39 independent samples and 262 measures of discounting, identified via a systematic review of PsychInfo, PubMed, and Google Scholar databases. We calculated omnibus effect-size estimates and evaluated the role of proposed moderators using a robust variance estimation meta-regression method. The meta-regression output reflected modest test-retest reliability, r = .670, p < .001, 95% CI [.618, .716]. Discounting was most reliable when measured in the context of temporal constraints, in adult respondents, when using money as a medium, and when reassessed within 1 month. Testing also suggested acceptable stability via nonsignificant and small changes in effect magnitude over time, dz = 0.048, p = .31, 95% CI [-0.051, 0.146]. Clinicians and researchers seeking to measure discounting can consider the contexts when reliability is maximized for specific cases.

Measuring the Dissemination Impact of Culturo-Behavioral Science.

Culturo-behavior science addresses many of the world's most significant problems and therefore has potential to create world-changing outcomes. Before systems level changes that improve the world can be implemented, however, it is first necessary for the public to know about and take interest in the accomplishments of culturo-behavior science. Measurable evidence that this kind of influence is being achieved is a component of "dissemination impact," an important but often overlooked form of accountability on sciences that target real-world problems. We describe a method for quantifying some aspects of dissemination impact and present data on how much of this impact has been earned by articles published in Behavior and Social Issues. The results provide considerable food for thought about how culturo-behavior science can proceed toward making the strategic pursuit of dissemination impact a component of its evidence-based practices. Supplementary Information: The online version contains supplementary material available at 10.1007/s42822-022-00120-3.