David J. Cox
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
David J. Cox has a total of 84 co-authors. Here are the top co-authors by number of shared publications:
- Matthew Thornton Brodhead
Michigan State University, East Lansing, USA
9 shared publications - Adrienne Michelle Jennings
Department of Behavioral Science, Daemen University, Amherst, USA
8 shared publications - Matthew W. Johnson
Center of Excellence for Psilocybin Research and Treatment , Sheppard Pratt
5 shared publications Asim Javed
Endicott College Beverly MA USA
4 shared publications- Jacob Sosine
Department of Data & Analytics; Behavioral Health Center of Excellence, Beverly, USA
4 shared publications - Jesse Dallery
Department of Psychology, University of Florida, Gainesville, USA
4 shared publications
Latest Publications
- Open Access
- 22 September 2026
- Published by Springer Nature in Journal of Behavioral Education
- Vol. 35 (3) , 1663-1664
Correction: Machine Learning to Analyze Alternating Treatments Graphs
- 22 September 2026
- Published by Springer Nature in Journal of Behavioral Education
- Vol. 35 (3) , 1663-1664
Open AccessCorrection: Machine Learning to Analyze Alternating Treatments Graphs
- 14 September 2026
- Published by Springer Nature in Behavior Analysis in Practice
- p. 1-16
Survey of Open Science Practices in Behavior Analysis
- 14 September 2026
- Published by Springer Nature in Behavior Analysis in Practice
- p. 1-16
Survey of Open Science Practices in Behavior Analysis
- 13 June 2026
- Published by Wiley in Behavioral Interventions
- Vol. 41 (3) , e70094
A Proof of Concept Demonstration for How to Evaluate Clinical Decision Models
Decision models are frequently published in applied behavior analysis (ABA) to guide clinical decision‐making, yet few have been empirically tested for their validity. To help researchers and practitioners identify potential weaknesses in decision models, the present study demonstrates one analytic method researchers can use to assess the procedural validity of decision models. Participants ( N = 75) were presented with systematically designed hypothetical vignettes to comprehensively test all decision pathways. Each vignette described a client, challenging behavior, and the targeted therapeutic environment (e.g., school, home) similar to what might be collected during an intake process. Participants were then asked to select the best data collection procedure for their initial behavioral observation either without (baseline) or with access to an associated decision model. Access to the decision model did not significantly improve performance for the control or behavior analyst group. Once these data were collected, we demonstrate how an error analysis allows researchers to identify which decision points incorrect decisions cluster at, providing information on how to improve the language of the decision model. Having an analytic technique that identifies challenges in decision model language and decision paths will aid future researchers in the iterative process i.e. decision model design and validation.
- 13 June 2026
- Published by Wiley in Behavioral Interventions
- Vol. 41 (3) , e70094
A Proof of Concept Demonstration for How to Evaluate Clinical Decision Models
Decision models are frequently published in applied behavior analysis (ABA) to guide clinical decision‐making, yet few have been empirically tested for their validity. To help researchers and practitioners identify potential weaknesses in decision models, the present study demonstrates one analytic method researchers can use to assess the procedural validity of decision models. Participants ( N = 75) were presented with systematically designed hypothetical vignettes to comprehensively test all decision pathways. Each vignette described a client, challenging behavior, and the targeted therapeutic environment (e.g., school, home) similar to what might be collected during an intake process. Participants were then asked to select the best data collection procedure for their initial behavioral observation either without (baseline) or with access to an associated decision model. Access to the decision model did not significantly improve performance for the control or behavior analyst group. Once these data were collected, we demonstrate how an error analysis allows researchers to identify which decision points incorrect decisions cluster at, providing information on how to improve the language of the decision model. Having an analytic technique that identifies challenges in decision model language and decision paths will aid future researchers in the iterative process i.e. decision model design and validation.
- 10 June 2026
- Published by Wiley in Journal of the Experimental Analysis of Behavior
- Vol. 126 (1) , e70118
Evaluating the combined effects of effort and probability on monetary discounting
Researchers have studied how the effort to obtain a reward affects its value and how effort discounting compares to other reinforcer characteristics (e.g., delay, social, probability). However, many everyday choices are likely influenced by how effort combines with other reinforcer characteristics, which have primarily been studied in isolation. In this experiment, we tested nine models describing how effort combines with probability to influence monetary discounting of $100 using an adjusting‐amount task. Overall, participants' ( N = 61) choices followed expected patterns: Steeper probability discounting was observed at higher effort levels, and steeper effort discounting was observed at lower probabilities. All nine effort‐probability models described group and individual data well. Using AICc, the general theory of discounting model best described most participants' data, with Rachlin's multiplicative hyperboloid model ranking second and a multiplicative combination of Rachlin's hyperboloid for probability and a power‐law for effort ranking third. In total, the results from model fitting indicate that probability and effort interact to influence choice. Existing models describe this interaction well, but future research will be needed to determine the conditions under which different models might be favored. These preliminary findings add effort‐probability interactions to the literature on the multiple control of choice.
- 10 June 2026
- Published by Wiley in Journal of the Experimental Analysis of Behavior
- Vol. 126 (1) , e70118
Evaluating the combined effects of effort and probability on monetary discounting
Researchers have studied how the effort to obtain a reward affects its value and how effort discounting compares to other reinforcer characteristics (e.g., delay, social, probability). However, many everyday choices are likely influenced by how effort combines with other reinforcer characteristics, which have primarily been studied in isolation. In this experiment, we tested nine models describing how effort combines with probability to influence monetary discounting of $100 using an adjusting‐amount task. Overall, participants' ( N = 61) choices followed expected patterns: Steeper probability discounting was observed at higher effort levels, and steeper effort discounting was observed at lower probabilities. All nine effort‐probability models described group and individual data well. Using AICc, the general theory of discounting model best described most participants' data, with Rachlin's multiplicative hyperboloid model ranking second and a multiplicative combination of Rachlin's hyperboloid for probability and a power‐law for effort ranking third. In total, the results from model fitting indicate that probability and effort interact to influence choice. Existing models describe this interaction well, but future research will be needed to determine the conditions under which different models might be favored. These preliminary findings add effort‐probability interactions to the literature on the multiple control of choice.
- 17 May 2026
- Published by Wiley in Journal of the Experimental Analysis of Behavior
- Vol. 125 (3) , e70109
Choice with ethical outcomes: Sign effect, commodity effect, and ethical theory preference
Researchers have found significant utility in approaching choice from a behavioral economics lens. Recently, this framework has begun to be extended to describe and predict clinical‐ethical choices. Researchers have found that common behavioral economics effects occur with clinical‐ethical choice at the group level (e.g., framing effects, effort discounting, delay discounting, and probability discounting) and the individual level (e.g., probability discounting with specific outcomes). However, past research examining clinical‐ethical outcomes through the behavioral economics lens has used experimental preparations that differ from those commonly used in discounting research with other commodities. Thus, the current exploratory study sought to use adjusting‐amount procedures commonly employed in discounting work to examine the sign effect in ethical probability discounting; the commodity effect between monetary and ethical outcomes; and correlations among rates of monetary discounting, ethical discounting, and ethical theory preference as measured via an exploratory, ethical multiple‐stimulus‐without‐replacement procedure. We observed monetary outcomes discounted more steeply than ethical outcomes, ethical violations discounted more steeply than ethical compliance, no significant correlations among discounting tasks, and no significant correlations between ethical theory preference and ethical discounting. The present work further demonstrates the presence of behavioral economics effects with ethical decision making. However, choice with ethical outcomes may not easily be predicted by choice with other commodities, measures in behavioral economics, or ethical theory preference. Despite these findings, many limitations remained, primarily centered on whether procedures common to behavioral economic analyses with other commodities are adequate for use when studying clinical‐ethical decision making. Moving forward, alternative preparations to study ethical decision making may be needed.
- 17 May 2026
- Published by Wiley in Journal of the Experimental Analysis of Behavior
- Vol. 125 (3) , e70109
Choice with ethical outcomes: Sign effect, commodity effect, and ethical theory preference
Researchers have found significant utility in approaching choice from a behavioral economics lens. Recently, this framework has begun to be extended to describe and predict clinical‐ethical choices. Researchers have found that common behavioral economics effects occur with clinical‐ethical choice at the group level (e.g., framing effects, effort discounting, delay discounting, and probability discounting) and the individual level (e.g., probability discounting with specific outcomes). However, past research examining clinical‐ethical outcomes through the behavioral economics lens has used experimental preparations that differ from those commonly used in discounting research with other commodities. Thus, the current exploratory study sought to use adjusting‐amount procedures commonly employed in discounting work to examine the sign effect in ethical probability discounting; the commodity effect between monetary and ethical outcomes; and correlations among rates of monetary discounting, ethical discounting, and ethical theory preference as measured via an exploratory, ethical multiple‐stimulus‐without‐replacement procedure. We observed monetary outcomes discounted more steeply than ethical outcomes, ethical violations discounted more steeply than ethical compliance, no significant correlations among discounting tasks, and no significant correlations between ethical theory preference and ethical discounting. The present work further demonstrates the presence of behavioral economics effects with ethical decision making. However, choice with ethical outcomes may not easily be predicted by choice with other commodities, measures in behavioral economics, or ethical theory preference. Despite these findings, many limitations remained, primarily centered on whether procedures common to behavioral economic analyses with other commodities are adequate for use when studying clinical‐ethical decision making. Moving forward, alternative preparations to study ethical decision making may be needed.
- Open Access
- 13 April 2026
- Published by Center for Open Science in PsyArXiv
Testing Behavior Dynamic Models to Predict Choice in a Non-Stationary Two-Choice Foraging Environment
Between the 1970s-1990s, researchers proposed several dynamic models of choice to explain how organisms adjust responding under changing contingencies (e.g., melioration, kinetic model, behavioral momentum, hill-climbing). Although widely debated, these models were rarely tested against one another on the same dataset, and alternative models from related sciences similarly predict individual choice based on contact with reinforcement. The present study revisits this literature by fitting 22 models to trial-level choice data from 60 humans completing a four-phase, non-stationary, two-option foraging task with regime shifts, contingency reversals, and perturbation probes. Models included three baseline heuristics, six historical models, three reinforcement-learning (RL) models, three hidden Markov models (HMMs), three foraging models, and four empirical dynamical models (EDM). Using AIC, HMMs with 3-4 latent states performed best (mean AIC = −2,010 to −4,183) followed by RL (≈ 1,040), historical models (≈ 1,052), then foraging models (≈1,216). Next-response predictive accuracy showed a similar pattern: HMMs performed best (≈ .88), followed by EDM (.74), RL (.69), historical (.66), foraging (.66) and baseline models (.53). Overall, state-switching architectures may better describe and predict choice in dynamic contexts where steady-state behavior is unlikely to emerge. Regardless, much work remains to predict response-by-response behavior dynamics across varying environmental contexts.
- 13 April 2026
- Published by Center for Open Science in PsyArXiv
Open AccessTesting Behavior Dynamic Models to Predict Choice in a Non-Stationary Two-Choice Foraging Environment
Between the 1970s-1990s, researchers proposed several dynamic models of choice to explain how organisms adjust responding under changing contingencies (e.g., melioration, kinetic model, behavioral momentum, hill-climbing). Although widely debated, these models were rarely tested against one another on the same dataset, and alternative models from related sciences similarly predict individual choice based on contact with reinforcement. The present study revisits this literature by fitting 22 models to trial-level choice data from 60 humans completing a four-phase, non-stationary, two-option foraging task with regime shifts, contingency reversals, and perturbation probes. Models included three baseline heuristics, six historical models, three reinforcement-learning (RL) models, three hidden Markov models (HMMs), three foraging models, and four empirical dynamical models (EDM). Using AIC, HMMs with 3-4 latent states performed best (mean AIC = −2,010 to −4,183) followed by RL (≈ 1,040), historical models (≈ 1,052), then foraging models (≈1,216). Next-response predictive accuracy showed a similar pattern: HMMs performed best (≈ .88), followed by EDM (.74), RL (.69), historical (.66), foraging (.66) and baseline models (.53). Overall, state-switching architectures may better describe and predict choice in dynamic contexts where steady-state behavior is unlikely to emerge. Regardless, much work remains to predict response-by-response behavior dynamics across varying environmental contexts.
- Open Access
- 25 March 2026
- Published by Center for Open Science in PsyArXiv
Toward a Dynamical Systems Account of Behavioral Networks: Integrating Matching, Disequilibrium Theory, Momentum, and Brunt-Väisälä Frequency
Behavior science has produced robust quantitative accounts of individual processes such as matching, disequilibrium, and behavioral momentum. However, research on these behavioral processes have largely developed in isolation from one another. This paper attempts a first step toward a unifying, dynamical systems framework that integrates these three traditions and extends them with an analogy drawn from fluid dynamics (i.e., the Brunt-Väisälä frequency). In the proposed framework, the matching law defines equilibrium geometry as the baseline allocation surface toward which behavior tends; disequilibrium theory provides the restoring vector field that arises when behavior is displaced from that surface; behavioral momentum supplies the inertial term governing resistance to change; and the Brunt-Väisälä analogy characterizes the rate and stability of return to equilibrium via reinforcer density gradients. Because contingencies operating on one response can propagate through an organism's repertoire, the framework further incorporates network coupling among responses. All components are combined into a single second-order differential equation governing behavior allocation over time. An implementation strategy based on supra-adjacency tensors and temporal-difference learning is described, and numerical simulations of an ABABABAB reversal design illustrate how the framework captures transitions between reinforcement conditions, the persistence conferred by reinforcement history, and the effects of inter-response coupling.
- 25 March 2026
- Published by Center for Open Science in PsyArXiv
Open AccessToward a Dynamical Systems Account of Behavioral Networks: Integrating Matching, Disequilibrium Theory, Momentum, and Brunt-Väisälä Frequency
Behavior science has produced robust quantitative accounts of individual processes such as matching, disequilibrium, and behavioral momentum. However, research on these behavioral processes have largely developed in isolation from one another. This paper attempts a first step toward a unifying, dynamical systems framework that integrates these three traditions and extends them with an analogy drawn from fluid dynamics (i.e., the Brunt-Väisälä frequency). In the proposed framework, the matching law defines equilibrium geometry as the baseline allocation surface toward which behavior tends; disequilibrium theory provides the restoring vector field that arises when behavior is displaced from that surface; behavioral momentum supplies the inertial term governing resistance to change; and the Brunt-Väisälä analogy characterizes the rate and stability of return to equilibrium via reinforcer density gradients. Because contingencies operating on one response can propagate through an organism's repertoire, the framework further incorporates network coupling among responses. All components are combined into a single second-order differential equation governing behavior allocation over time. An implementation strategy based on supra-adjacency tensors and temporal-difference learning is described, and numerical simulations of an ABABABAB reversal design illustrate how the framework captures transitions between reinforcement conditions, the persistence conferred by reinforcement history, and the effects of inter-response coupling.
- 12 February 2026
- Published by Springer Nature in Behavior Analysis in Practice
A Guide to Choosing a Balanced Outcomes Measurement Portfolio for ABA Organizations
- 12 February 2026
- Published by Springer Nature in Behavior Analysis in Practice
A Guide to Choosing a Balanced Outcomes Measurement Portfolio for ABA Organizations