Translated from the Portuguese original. Read the original (Português)
Behavioral momentum is a metaphor borrowed from physics to study the persistence of behavior. The rate at which a behavior occurs works like the velocity of a moving body; how much it resists when something disrupts it works like that body's mass. And the two things don't necessarily go together: studies with pigeons and with adults show that what makes a behavior hard to change is the amount of reinforcement available in that context — including reinforcement that goes to other behaviors. Hence an uncomfortable consequence: a treatment that works can, without anyone noticing, leave the problem behavior more persistent than before.
Behavioral momentum theory is a quantitative model that treats a response's persistence as a property distinct from its rate. In the formulation by Nevin and Grace (2000), the rate of a free operant in the presence of a discriminative stimulus is analogous to the velocity of a moving body, and resistance to change — how much the response is maintained, relative to its own baseline, when something disrupts it — measures an aspect analogous to that body's inertial mass. In an extension of the metaphor, preference measures an analogue to gravitational mass. The independent functions relating resistance to change and preference to reinforcement conditions can be read as convergent measures of a single construct, analogous to physical mass, which represents the effects of a history of exposure to signaled conditions of reinforcement.
A metaphor borrowed from physics
In mechanics, the momentum (or "quantity of motion") of a body is the product of its mass and its velocity. Two balls can roll at the same speed and still, one of them be much harder to stop — because it's heavier. The question behavioral momentum theory asks is exactly that, transposed to behavior: besides knowing how often a behavior happens, how do we know how much it will hold up when the environment changes?
The answer proposed by John Nevin, in 1974, was to treat resistance to change as the measure of a response's "strength." Not the rate: the resistance. In 2000, Nevin and Grace organized this idea as a modern, quantitative version of Thorndike's Law of Effect (1911) — with a practical implication that runs through this entry: rate and persistence can vary in opposite directions.
How persistence is measured
Measuring persistence requires comparing two behaviors under the same disruption. The classic arrangement for this is the multiple schedule: two or more reinforcement schedules that alternate, each signaled by its own discriminative stimulus — this is how the field's reference manuals, such as Cooper, Heron, and Heward's, define the arrangement. Once the two responses stabilize, a disruptor comes into play: something that disrupts both at the same time. It can be extinction, satiation (for example, giving access to food before the session), a distraction, or an outside reinforcer. What gets compared, then, is not the raw rates, but the proportion that each response retains of its own baseline.
In the study that launched this line of reading, Nevin (1974) trained pigeons to peck in the presence of one key color under one reinforcement schedule or condition, and in the presence of another color under another schedule or condition. Once performance stabilized, he began presenting response-independent food during the intervals between components, with the key dark. Rates dropped in both components — but the drop relative to baseline depended on the reinforcement conditions in force in that component: the frequency, the magnitude, the delay, or the requirement of a certain response rate. When reinforcement was ended for good, the drops ordered themselves the same way as under response-independent food. The relations between the components' rates were power functions. Nevin's conclusion was that the concept of response strength summarizes the effects of very different procedures — and that this strength is identified with relative resistance to change. That was the initial formulation of the idea; the next step was discovering what, exactly, builds this resistance.
Rate and persistence are not the same thing
One of the field's most counterintuitive results came from an experiment that deliberately separated the two measures. Nevin, Tota, Torquato, and Shull (1990) reinforced pigeons' pecks with variable-interval schedules in two components and, in some conditions, added food delivered independently of the response in one of the components. What happened was a clear dissociation:
- The rate of pecking in the component got lower the greater the proportion of reinforcers coming from the alternative source — in other words, the "free" food reduced the measured behavior.
- The resistance to satiation and extinction, in contrast, got higher as the overall reinforcement rate of that component rose — even with a lower response rate.
In a second experiment, the alternative reinforcers became contingent on a specific concurrent response, and the pattern repeated. The authors' conclusion is the backbone of the theory: the response's relative reinforcement — the operant contingency — determines its rate; the stimulus–reinforcer relation — a Pavlovian contingency, the kind studied in respondent conditioning — determines its resistance to change. In other words: what makes a behavior persistent is the richness of the context in which it was learned, not just what the behavior itself produces.
Outside the laboratory
The theory was born with pigeons, but it has been tested in far less controlled situations. Explore three published applications:
Popcorn, coffee, and a video playing
In Mace and colleagues (1990), adults with intellectual disabilities who lived in a group home received popcorn or coffee for sorting plastic utensils. Under a multiple schedule, sorting utensils of one color was reinforced on average every 60 seconds; those of another color, on average every 240 seconds. The baseline rates were similar — but when a video program started playing at the same time, the task resisted the distraction more precisely in the color correlated with the higher reinforcement rate. In the second part, the two colors started paying equally and one of them received extra reinforcers independent of the task: there the baseline rate got lower and resistance to distraction, higher. Resisting distraction depended on the reinforcers obtained in the presence of the stimulus, not on response rates or on the contingencies between response and reinforcer.
Reinforcer quality and following instructions
The high-probability instructional sequence is a well-known procedure for dealing with refusal to follow requests: you first ask for something the person usually does easily, then present the difficult request (the sequence has a dedicated page in our Evidence-Based Practice section). It works — but it also fails sometimes. Mace, Mauro, Boyajian, and Eckert (1997) tested the hypothesis that using higher-quality reinforcers for the easy instructions would improve the procedure, in two applied studies and one basic laboratory study, the latter with two different solutions as reinforcers and an extinction test. The three studies showed an overall consistent relation between reinforcer quality and behavioral momentum — preference assessment thus gains a role that goes beyond "motivating."
Why the coach calls a timeout
Mace and Lalli (1992) recorded, from videos of 14 games of the 1989 U.S. college tournament, three classes of events: reinforcers (such as points and turnovers won from opponent errors), adversities (such as missed shots, turnovers, and fouls), and the team's response to adversity (the outcome of the next possession). Three findings: the favorable response to adversity generally increased as the reinforcement rate in the previous 3 minutes rose; coaches called timeouts when they were behind on the scoreboard, on average, by 2.63 to 1; and calling a timeout seemed to work as an intervention to reduce the opponent's reinforcement rate — in the 3 minutes following the timeout, the two teams' rates became nearly equal. These are descriptive data, discussed in light of behavioral momentum.
Why relapse happens
Here the theory stops being a laboratory curiosity. According to Nevin and Shahan (2011), all reinforcers obtained in the presence of a discriminative stimulus increase the resistance to change of the behavior that occurs there — regardless of whether they are contingent on the target behavior, delivered with no contingency at all, or contingent on an alternative behavior. This claim is where the uncomfortable prediction comes from: differential reinforcement of an alternative behavior (DRA) — one of the best-supported interventions for challenging behaviors — enriches the context. It reduces the frequency of the problem behavior and, at the same time, can increase its persistence.
When reinforcement of the alternative is interrupted — by a school change, by the team's vacation, by an atypical day —, the old behavior can reappear. This effect has a name: resurgence. In the model by Shahan and Sweeney (2011), alternative reinforcement during extinction of the target acts in two ways at once: as an additional source of disruption, which helps knock down the problem behavior in the present, and as a source of reinforcement in the context, which increases its future strength. As Pritchard, Hoerger, and Mace (2014) note in a review of the field, relapse after an apparently successful treatment is a persistent problem in applied behavior analysis — and momentum theory offers a way to predict it rather than just lament it.
The study that tested a way outMace and colleagues (2010) conducted three coordinated experiments. In the first, a clinical one, resistance to extinction was substantially higher after high DRA reinforcement rates than after baseline rates — confirming the unwanted effect. In the second, a rat model, training the alternative response in a context without reinforcement of the target response circumvented this strengthening of persistence. In the third, back in the clinic, applying the DRA in a separate context produced less resistance to extinction than applying it in the same context where the target behavior was reinforced. The practical lesson is not to abandon the DRA: it's to pay attention to where it is taught.
There is also evidence that the persistence of the desired behavior can be built on purpose. Wacker and colleagues (2011) followed eight young children whose destructive behavior was maintained, at least in part, by negative reinforcement — escaping or getting out of a task. They received long-term functional communication training: they completed part of the task and touched a communication card to get a brief break. Before and throughout treatment, extinction probes tested the relative persistence of the appropriate behavior and the destructive one. Treatment often reduced the destructive behavior quickly — and yet, in probes done early in treatment, the destructive behavior tended to come back. But after the treatment's effects started to persist in these probes, four subsequent challenges (longer extinction sessions, a new task, withdrawal of the communication card, and a schedule in which both behaviors were reinforced) produced only mild to moderate disruptions. The authors suggest this offers an alternative definition of maintenance: maintenance as behavioral persistence.
Three terms come up often when discussing this topic. Tap the cards:
What the theory still hasn't resolved
Behavioral momentum theory organized decades of data and changed the conversation about relapse — but it is not a closed subject, and it's the authors themselves who say so. In a review with the good-humored subtitle "fits and misfits," Nevin, Craig, Cunningham, Podlesnik, Shahan, and Sweeney (2017) suggest the theory's extension can account for other forms of relapse — those studied under the names reinstatement and renewal — but that its application to resurgence is flawed — both conceptually and because it fails to explain more recent data. They propose that the increase in persistence produced by alternative reinforcers is limited to their concurrent availability within a distinctive stimulus context, and they acknowledge that alternatives are needed that can account for resurgence without losing what the theory already explains well. They also record a result that troubles the Pavlovian explanation itself: in a Pavlovian-instrumental transfer paradigm — which tests the effect of a Pavlovian stimulus on an operant response —, the expected effects of reinforcer rates correlated with the stimulus were not found.
Another quantitative model, studied alongside momentum theory, is resurgence as choice. In a brief 2025 review, Kranak and colleagues observe that both models could improve treatments if used prospectively — that is, to plan the intervention before relapse happens — but that it is unclear how often this actually occurs in practice, and they discuss the barriers to it happening.
Three common misunderstandings1. It is not willpower or "habit inertia." The mass in the metaphor is not a property of the person: it's the summary of a reinforcement history in that context. Changing the context changes persistence. 2. Doing a lot is not the same as resisting a lot. Nevin, Tota, Torquato, and Shull (1990) and Mace and colleagues (1990) added response-independent reinforcement and saw the rate drop while resistance rose. A high rate doesn't guarantee persistence, and a low rate doesn't indicate fragility. 3. The prediction about DRA is not an argument against treating. Differential reinforcement remains among the best-supported interventions for challenging behaviors; what the theory adds is a planning alert — about the context in which the alternative is taught and about what to do when treatment is interrupted. None of this replaces a functional analysis and follow-up by a qualified professional.
Test your understanding
Quick quiz
Three questions to lock in the concept. You'll see the explanation right away.
Further reading
- Nevin, J. A. “Response strength in multiple schedules”. Journal of the Experimental Analysis of Behavior, 21(3), 389–408, 1974.
- Nevin, J. A., Tota, M. E., Torquato, R. D. & Shull, R. L. “Alternative reinforcement increases resistance to change: Pavlovian or operant contingencies?”. Journal of the Experimental Analysis of Behavior, 53(3), 359–379, 1990.
- Mace, F. C., Lalli, J. S., Shea, M. C., Lalli, E. P., West, B. J., Roberts, M. & Nevin, J. A. “The momentum of human behavior in a natural setting”. Journal of the Experimental Analysis of Behavior, 54(3), 163–172, 1990.
- Mace, F. C. & Lalli, J. S. “Behavioral momentum in college basketball”. Journal of Applied Behavior Analysis, 25(3), 657–663, 1992.
- Mace, F. C., Mauro, B. C., Boyajian, A. E. & Eckert, T. L. “Effects of reinforcer quality on behavioral momentum: coordinated applied and basic research”. Journal of Applied Behavior Analysis, 30(1), 1–20, 1997.
- Nevin, J. A. & Grace, R. C. “Behavioral momentum and the law of effect”. Behavioral and Brain Sciences, 23(1), 73–90, 2000.
- Mace, F. C., McComas, J. J., Mauro, B. C., Progar, P. R., Taylor, B., Ervin, R. & Zangrillo, A. N. “Differential reinforcement of alternative behavior increases resistance to extinction: clinical demonstration, animal modeling, and clinical test of one solution”. Journal of the Experimental Analysis of Behavior, 93(3), 349–367, 2010.
- Shahan, T. A. & Sweeney, M. M. “A model of resurgence based on behavioral momentum theory”. Journal of the Experimental Analysis of Behavior, 95(1), 91–108, 2011.
- Nevin, J. A. & Shahan, T. A. “Behavioral momentum theory: equations and applications”. Journal of Applied Behavior Analysis, 44(4), 877–895, 2011.
- Wacker, D. P., Harding, J. W., Berg, W. K., Lee, J. F., Schieltz, K. M., Padilla, Y. C., Nevin, J. A. & Shahan, T. A. “An evaluation of persistence of treatment effects during long-term treatment of destructive behavior”. Journal of the Experimental Analysis of Behavior, 96(2), 261–282, 2011.
- Pritchard, D., Hoerger, M. & Mace, F. C. “Treatment relapse and behavioral momentum theory”. Journal of Applied Behavior Analysis, 47(4), 814–833, 2014.
- Nevin, J. A., Craig, A. R., Cunningham, P. J., Podlesnik, C. A., Shahan, T. A. & Sweeney, M. M. “Quantitative models of persistence and relapse from the perspective of behavioral momentum theory: fits and misfits”. Behavioural Processes, 141(Pt 1), 92–99, 2017.
- Kranak, M. P., Falligant, J. M., Jones, C., Stephens, M. & Wessel, M. “On the prospective application of behavioral momentum theory and resurgence as choice in the treatment of problem behavior: a brief review”. Behavioral Sciences, 15(5), 688, 2025.
- Cooper, J. O., Heron, T. E. & Heward, W. L. Applied Behavior Analysis. 3rd ed. Pearson, 2020.
This entry is educational and does not replace assessment, diagnosis or treatment by a qualified health professional. Every person is unique — for a specific case, seek individual guidance.