Join Free →
About Weekly CEU Resource Hubs ↳ CEU Library ↳ Research Library Blog Contact Join Free →
Topic Guide · Practitioner

The Premack Principle: A Practitioner's Reference for BCBAs, RBTs, and School Behavior Teams

Query target: Premack principle · BBC Editorial Team
★ Summary

The Premack principle, formulated by David Premack (1959, 1965), states that a higher-probability behavior — measured by free-operant baseline rate — can reinforce a lower-probability behavior when contingent access to the high-probability behavior is made dependent on the low-probability one. In day-to-day BCBA, RBT, and classroom practice this is "Grandma's rule" — first work, then play — and it sits beneath every visual schedule, work-then-break ratio, and contingent-access reinforcement plan in applied behavior analysis (BACB Task List, 5e B-4). The principle is foundational, but the modern empirical refinement — Timberlake and Allison's response deprivation hypothesis — is what actually predicts when contingent access will function as reinforcement. This page is a working reference for what the principle says, what its successors say, and how to apply both correctly. Note that Premack's original 1959 and 1965 papers predate the open-access era and are not in this corpus; we cite them in plain text and lean on contemporary work that operationalizes the principle in modern ABA (Premack, 1959; Premack, 1965).

01What the Research Says

Premack's original formulation

Premack's 1959 paper in Psychological Review and the more developed 1965 chapter introduced what he called the probability-differential hypothesis: behaviors are arrayed on a relative-rate hierarchy, and any higher-rate behavior on that hierarchy can reinforce any lower-rate behavior contingently (Premack, 1959; Premack, 1965). Classic demonstrations involved rats (running vs. drinking) and cebus monkeys; the logic was later extended to children's free-play behavior. Two pieces are critical for practitioners. First, probability is measured under free-operant baseline, not from preference rankings or topography. Second, the relation is reversible: deprive an animal of the lower-probability behavior long enough and the hierarchy flips (Premack, 1965).

This was a real conceptual break from Skinner. In Skinner's transitive framework a stimulus is a reinforcer based on its history — the predicate is attached to the stimulus. In Premack's framework, reinforcement is a relation between two responses at a given moment, defined by their relative rates. That distinction is why the Premack principle still matters when token economies and DRA fall short: when nothing on a preference hierarchy is currently functioning as a reinforcer, you can sometimes find a reinforcing relation by looking at what the learner is already doing.

From Premack to motivating operations

Jack Michael's later work folded Premack's probability-differential insight into the broader concept of the motivating operation (MO), the antecedent variable that alters the value of a consequence and the frequency of behavior associated with it (Esch & Esch, 2016). Klatt and Morris (2001) explicitly drew the line linking Premack's principle, response deprivation, and establishing operations (Klatt & Morris, 2001). For modern practitioners, this matters because the question "is X a reinforcer?" is rarely answered by topography or even by a preference assessment — it is answered by the current value of contingent access to X relative to the learner's current level of engagement with X. That is a motivating-operation question, and Premack is its empirical ancestor (BACB Task List, 5e B-12).

Response deprivation: the operational replacement

By the late 1970s, Timberlake and Allison's response deprivation hypothesis had emerged as a more empirically robust framework than the strict Premack formulation. Their argument: contingent access to a behavior functions as reinforcement whenever the contingency restricts the responder below baseline rate, regardless of whether the contingent behavior is higher- or lower-probability than the instrumental one. This explains cases the original Premack principle cannot — for example, when a lower-probability behavior reinforces a higher-probability behavior, which response deprivation predicts whenever the schedule deprives the learner of the lower-rate behavior relative to baseline. Practically, a 5-minute drawing block can reinforce a 30-second math task even if math has the higher free-operant rate, provided the contingency restricts drawing below baseline. Jacobs and colleagues' tutorial on the disequilibrium model ties Premack, Killeen's mathematical reformulation (Killeen, 2014), and response deprivation into one predictive framework (Jacobs et al., 2017). For BCBAs writing behavior plans in 2026, response deprivation/disequilibrium is the framework to think in; Premack is the historical anchor that explains why it works.

Premack as a clinical lever for mand training

The most direct contemporary application of the Premack principle in the corpus is in mand instruction. Frampton and colleagues' clinical tutorial on capturing and contriving establishing operations builds the entire procedure around scheduling preferred (high-probability) activities as immediate consequences for mand responses, with temporary restriction of access used to re-establish the value-altering effect when preferred items lose evocative power (Frampton et al., 2024). The procedural anchor is straight Premack: low-probability behavior (the mand response) gains strength because contingent access to high-probability behavior (engaging with the preferred item) is made dependent on it (Frampton et al., 2024).

Reinforcer density and rule-governed choice

A laboratory analogue of the Premack principle in adults appears in Ruiz Méndez's concurrent-schedule rule-governed-choice procedure, where six of eight adults preferred the verbally instructed alternative correlated with the richer (6 reinforcers/min) over leaner (1 reinforcer/min) schedule when both rules were simultaneously available (Méndez, 2024). The result is consistent with the Premack-style claim that relative reinforcement density controls choice — and it argues for a practical rule when designing contingency-based interventions: pair the target response with the richest available source of reinforcement and use clear discriminative stimuli to mark when the contingency is in effect (Méndez, 2024).

Where the strict Premack formulation breaks down

The principle has known limits. Behavior is sensitive to relative rather than absolute reinforcer value, and contrast effects can flip the hierarchy in ways the strict Premack formulation does not predict — de Haan and Simon's reformulation uses the classic monkey-refusing-lettuce-after-bananas demonstration to argue that organisms discriminate relative reinforcer value rather than being mechanically pushed by a single response-strength variable (de Haan & Simon, 2024). The practical implication for BCBAs: when a high-value reinforcer immediately precedes a low-value Premack contingency, expect a contrast-induced performance drop. Sequencing matters. Preference instability — the routine reality that a child's "high-probability" activity at 8 a.m. may not be high-probability at 11 a.m. — is the same problem in clinical clothing.

02Evidence Tier Breakdown

The foundational Premack (1959, 1965) and Timberlake & Allison (1974) papers predate this corpus and are cited here in plain text (Premack, 1959; Premack, 1965). The contemporary research synthesized above sits primarily at the single-subject tier (Frampton et al., 2024) and the theoretical / tutorial tier (Jacobs et al., 2017), with no head-to-head comparative-effectiveness trial of Premack-only versus response-deprivation-framed contingencies in this corpus (Premack, 1965). The convergent operational claim — that contingent access to a higher-rate behavior reinforces a lower-rate one, with response deprivation as the more empirically robust modern framework — is well-supported across the SCED and theoretical literature (Jacobs et al., 2017).

03Decision Logic

When the Premack principle is the right tool, when response deprivation/disequilibrium is the better tool, and when neither is enough:

  1. Use Premack contingencies (first-then) when baseline observation shows a clear high-probability behavior the learner reliably engages with under free-operant conditions, the low-probability target behavior is a known operant (not a skill deficit), and the ratio of work-to-access is small enough that satiation is unlikely (Frampton et al., 2024).
  2. Move to response deprivation/disequilibrium framing when the contingent activity is not the highest-probability behavior in the repertoire, or the contingency has to restrict baseline engagement to function — the disequilibrium model handles this directly while strict Premack does not (Jacobs et al., 2017).
  3. Layer in a token economy or DRA when the work-to-access ratio for raw Premack would require unrealistically long instructional blocks before access, when multiple target behaviors need simultaneous reinforcement, or when the natural high-probability behavior cannot be safely or practically embedded in the setting (most school classrooms).
  4. Reach for a richer differential-reinforcement protocol (DRA, DRO, FCT) when the target behavior is challenging behavior with an identified function — Premack alone is not function-based, and pairing it with an unmatched reinforcer can fail or backfire.
  5. Re-baseline preferences whenever the contingency stops working — the most common implementation failure is assuming a previously high-probability activity is still high-probability today (BACB Task List, 5e B-12).

04Across Settings

Classroom (K–12)

Premack contingencies are the operational backbone of first-then boards, work-then-break ratios, and "complete five problems, earn three minutes of choice." In practice, the work-to-access ratio is the variable teachers most often miscalibrate — if the work block is too long relative to the access block, the contingency stops functioning and the student stops responding. Visual completion cues (a check card, a token strip, a timer) make the contingency discriminable, which matters for reinforcer-density effects on rule-following (Méndez, 2024). Rotating the access menu prevents satiation; embedding the same contingency across routines (circle time, centers, transitions) supports generalization.

Home and parent-implemented programming

At home, the Premack principle is the most parent-trainable contingency in ABA. Homework, then screen and clean up, then snack are immediately understandable and require no token system. The BCBA's role is calibrating ratios (parents systematically over-demand before access), rotating access activities to protect against satiation, and writing down the exact low-probability behavior the contingency targets so caregivers do not silently expand it.

Clinic and structured 1:1 instruction

In clinic settings, Premack contingencies appear explicitly in mand training, where preferred activities are scheduled as immediate consequences for mand responses and access is briefly restricted to re-establish evocative power (Frampton et al., 2024). They also serve quietly as social-validity reinforcers — a child who accepts a target bite during feeding therapy gets a preferred bite next, for example. The pattern to watch is satiation: a preferred activity run for 90 minutes loses its Premack value by 45.

Self-management and adult clients

For older learners and adults, Premack contingencies support self-management protocols: scheduling a low-probability task (a vocational chore, a fitness routine) immediately before a high-probability one (a meal, social media access), with the learner setting the ratio. Disequilibrium framing matters here because adult clients often have lower-probability activities that still function as reinforcers when the schedule deprives them below baseline (Jacobs et al., 2017).

05Common Pitfalls

  • Assuming preferred = reinforcing. A preference assessment ranks stimuli; it does not measure free-operant rate or current MO state. A "high-preference" item that the learner has had unrestricted access to all morning is no longer high-probability. Always corroborate preference data with current contingent-access function (BACB Task List, 5e B-12).
  • Mis-tuning the work-to-access ratio. The classic teacher error: 30 minutes of math, 2 minutes of break. Premack contingencies fail when the work requirement satiates the learner before access is delivered. Start small (1:1 or 2:1 work-to-access by time) and thin systematically.
  • Ignoring satiation across the day. A high-probability activity at 9 a.m. is often not high-probability at 2 p.m. Rotate access activities and re-baseline preferences whenever the contingency stops controlling behavior.
  • Confusing Premack with function-based intervention. Premack contingencies are not matched to the function of challenging behavior. Pairing Premack with an extinction component for a behavior maintained by escape can fail badly if the contingent reinforcer is not itself a functional match.
  • Forgetting response deprivation when the "high-probability" activity isn't actually highest. When the contingency works in spite of the learner having a clearly more-preferred alternative, the operative variable is response deprivation, not Premack proper (Jacobs et al., 2017).
  • Allowing post-reinforcer contrast to wreck the next demand. Sequencing a high-value reinforcer immediately before another demand without a transition often produces a contrast-induced performance drop on the next task (de Haan & Simon, 2024). Build a clear transition signal between access and the next demand.

06Practitioner Takeaways

  1. Premack defines reinforcement as a relation between two behaviors, not a property of a stimulus. The contingency is first low-probability, then high-probability — measured by current free-operant rate, not by preference ranking or topography (Premack, 1959; Premack, 1965).
  2. In modern ABA, write contingencies in disequilibrium/response-deprivation language even when you teach the concept as "Premack." The empirical predictions hold up in more cases (Jacobs et al., 2017).
  3. Always baseline the "high-probability" behavior under free-operant conditions before assuming it will reinforce. Preferred is not the same as high-probability, and high-probability now is not the same as high-probability after 30 minutes of access (BACB Task List, 5e B-12).
  4. Calibrate the work-to-access ratio small and thin systematically. Most Premack contingency failures in classrooms are ratio failures, not principle failures.
  5. Rotate the access menu to defeat satiation. A single high-probability activity used as the universal reinforcer across a 4-hour clinic session will stop being high-probability by hour two.
  6. Use Premack as the operational backbone of mand training. Schedule preferred activity access immediately contingent on the mand, restrict access briefly to re-establish evocative power, and probe across vocal/sign/AAC modalities under captured EOs before contriving them (Frampton et al., 2024).
  7. Pair the target response with the richest available reinforcement source and signal the contingency clearly. Reinforcement-density effects control rule-following when both rules are concurrently available (Méndez, 2024).
  8. Plan for contrast effects when sequencing reinforcers. A high-value reinforcer followed immediately by another demand often produces a performance drop on that next demand; build a transition signal (de Haan & Simon, 2024).
  9. Do not treat Premack as a function-based intervention for challenging behavior. Pair it with DRA, FCT, or function-matched extinction when the target is problem behavior, not skill acquisition.
  10. Document the exact low-probability behavior, the exact contingent activity, and the exact ratio. Categorical agreement that "we use first-then" hides specific-feature drift across staff and caregivers; the BIP should script the precise contingency in a way a second clinician could implement without you.

07Frequently Asked Questions

What is the Premack principle in plain language?

The Premack principle says that a behavior the learner is already doing at a higher rate (under free-operant baseline) can reinforce a behavior the learner is doing at a lower rate, when access to the high-rate behavior is made contingent on doing the low-rate one. The everyday version is "Grandma's rule" — first work, then play. Formally, reinforcement is defined as a relation between two responses, not as a property of a stimulus (Premack, 1959; Premack, 1965).

How is the Premack principle different from regular reinforcement?

In Skinner's transitive framework a stimulus is a reinforcer based on its conditioning history — the predicate is attached to the stimulus. In Premack's framework, reinforcement is a relation between two behaviors at the current moment, defined by their relative free-operant rates. Practically, this means a stimulus you've called a reinforcer can stop functioning if engagement with it under free-operant conditions drops below the engagement level of your "target" behavior, and a stimulus you haven't conditioned can suddenly start functioning if the schedule deprives the learner of it (BACB Task List, 5e B-4).

What is the response deprivation hypothesis, and why do practitioners care?

Timberlake and Allison's hypothesis says contingent access to a behavior functions as reinforcement whenever the contingency restricts the responder below baseline rate — regardless of whether the contingent behavior is higher- or lower-probability than the instrumental behavior. It is more empirically robust than strict Premack because it predicts cases where lower-probability behaviors reinforce higher-probability ones, which Premack alone cannot. The disequilibrium model unifies both into one predictive framework (Jacobs et al., 2017).

When should I choose a token economy over a Premack contingency?

Choose tokens when (a) the work-to-access ratio for raw Premack would require unrealistically long instructional blocks, (b) you need to reinforce multiple target behaviors simultaneously, (c) the natural high-probability behavior cannot be safely embedded in the setting, or (d) you need to bridge delays between target behavior and access. Tokens are not better than Premack — they are better at scaling and bridging delay. Both rest on the same underlying contingency logic (BACB Task List, 5e B-4, B-5).

Does the Premack principle work for challenging behavior?

Not directly. Premack contingencies build the acquisition side of programming — they reinforce a behavior you want to increase. For challenging behavior maintained by an identified function (escape, attention, tangible, automatic), the right tools are differential reinforcement (DRA/DRO) and function-based interventions like FCT, where the reinforcer is matched to the function. Premack can be a component inside a broader plan — for example, contingent access to a preferred activity following a functional communication response — but it is not itself function-based.

Why does my Premack contingency stop working after a few weeks?

Two common reasons. First, satiation: the previously high-probability activity is no longer high-probability under current free-operant conditions. Rotate the access menu and re-baseline preferences. Second, ratio drift: staff or caregivers gradually expand the work requirement before access ("just one more problem") until the ratio is no longer reinforcing. Document the exact ratio in the BIP and re-train fidelity periodically (BACB Task List, 5e B-12).

Is the Premack principle still considered current best practice in 2026?

Yes, with a refinement. The principle is universally taught and implemented in ABA, and appears in BACB Task List 5e content on positive reinforcement and motivating operations (BACB Task List, 5e B-4, B-12). For predicting whether a Premack contingency will function in a given case, the response deprivation hypothesis and the broader disequilibrium model are the more empirically robust frameworks, and current practitioner-facing tutorials embed Premack inside that wider account (Jacobs et al., 2017).

08References

Foundational sources predate the open-access era and are listed in plain text; modern ABA papers are linked to their BBC corpus pages above.