Sleeper Effect¶

Explain why a message paired with a discounting cue can grow more persuasive over weeks: the persuasive content and the cue that braked it are stored as separate memory traces that decay at different rates, so when the cue fades faster the net persuasion inverts sign at delayed recall.

Core Idea¶

The sleeper effect is the persuasion-psychological phenomenon in which a message delivered together with a discounting cue — a low-credibility source, a counter-argument, a disclaimer — produces a small or negligible immediate attitude change, but then shows increasing attitude change in the message's direction over days to weeks, the opposite of the usual pattern in which persuasion decays over time. The growing delayed effect is produced by the differential decay rates of two separable memory traces: the persuasive content of the message persists while the discounting cue — the source attribution, the caveat, the counter-argument — fades faster, so what survives at delayed retrieval is the message's persuasive force stripped of the brake that initially suppressed it.

The mechanism was identified in wartime research by Hovland and colleagues: audiences shown documentary films paired with a discounting source label showed small attitude shifts at immediate measurement but larger shifts in the film's direction when re-surveyed weeks later. Crucially, re-presenting the discounting cue at the time of the delayed survey eliminated the effect — the attitude reverted — confirming that the sleeper pattern depends specifically on the cue's absence from the retrieval environment, not on any change in how the content itself was remembered. The structural commitment is therefore a source-monitoring dissociation over time: content memory and cue memory are encoded separately, decay at different rates, and integrate at retrieval such that the delayed net persuasion is the content's push minus however much of the cue's brake remains — with a sign reversal relative to the immediate measurement whenever the cue has decayed sufficiently faster than the content.

Structural Signature¶

Sig role-phrases:

the discounted message — a persuasive communication delivered together with a discounting cue (low-credibility source, counter-argument, disclaimer)
the content trace — the memory of what was said, the persuasive force, encoded separately
the cue trace — the memory of who said it or why to doubt it, the brake, encoded separately from the content
the differential decay — the two traces fade at different rates, with the cue trace fading faster than the content trace
the retrieval integration — at recall the net persuasion is the content's push minus whatever of the cue's brake survives
the sign reversal — when the cue has decayed sufficiently faster, delayed attitude change grows in the message's direction, inverting the small immediate effect (opposite of ordinary decay)
the retrieval-environment locus — the effect lives in the cue's absence at recall, not in any change to how the content is remembered
the re-pairing test — re-presenting the cue at the delayed measurement abolishes the effect and reverts the attitude, isolating the locus and inverting the naive lever (make the cue persist, not the content less memorable)

What It Is Not¶

Not a message that simply "didn't work." A small immediate effect can be exactly backwards. Where the message arrived with a discounting cue, persuasion can grow in the message's direction over days to weeks as the cue fades faster than the content — so a campaign read as inert at exposure must be re-measured at delay, because its immediate figure may invert.
Not a change in how the content is remembered. The delayed gain lives in the retrieval environment — the cue's absence at recall — not in any strengthening of the content trace. The re-pairing test proves it: re-present the discounting cue at the delayed measurement and the attitude reverts, because the brake is back.
Not ordinary persuasion decay. Most messages fade toward baseline as memory weakens; the sleeper effect is a sign reversal — net persuasion grows past the immediate figure. It is the opposite of the usual time course, produced by the cue trace decaying faster than the content trace.
Not the mere exposure effect. Mere exposure raises liking through repetition, independent of content. The sleeper effect is specifically discounted content recovering its persuasive force as the discount fades — a cue-content dissociation over time, not a familiarity boost from repeated contact.
Not an automatic or guaranteed effect. It requires three preconditions: a discounting cue delivered with the message, a cue trace that decays faster than the content, and a retrieval environment lacking the cue. Absent any of these the sign does not reverse — which is why the effect is historically fragile and only appears under the specific conditions that let the brake fade independently.
Not the misinformation effect. Both involve content surviving while its source or discounting context is lost — the same source-versus-content memory architecture — but the sleeper effect is the persuasion/attitude version: a rejected claim regaining force. The misinformation effect concerns post-event contamination of memory for an event, a different task.

Scope of Application¶

The sleeper effect lives across persuasion and memory psychology, in the settings that rest on the same human source-versus-content memory architecture — separable cue and content traces decaying at different rates and integrating at retrieval; its reach stays within that substrate, since the substrate-neutral "two coupled traces fade at different rates, so their net effect inverts" lesson belongs to the (still emergent) differential_decay / cue-content-dissociation pattern, not to this name.

Persuasion research — the home turf: the Hovland wartime-documentary-film studies and decades of replications, the standard target case.
Advertising — the long-arc fade of disclaimers and source-credibility warnings, so consumers end up persuaded by claims they explicitly rejected at exposure.
Health messaging — a risk warning paired with appealing content fading while the appeal persists.
Eyewitness memory and misinformation — the near-relative source-monitoring failure, where a claim survives while the evidentiary context that should have discounted it is lost (the same memory architecture in a different task).

Clarity¶

Naming the sleeper effect splits a quantity that persuasion practice routinely treats as one: immediate persuasion and delayed persuasion become separable measurements with different dynamics rather than a single number sampled at one moment. Without the label, a message that produces little attitude change at exposure reads as simply ineffective — "it didn't work" — and the analyst stops measuring. The effect tells them that reading can be exactly backwards: where the message arrived with a discounting cue, the persuasive content and the brake on it are stored as separate memory traces that decay at different rates, so a campaign that looks inert at exposure can grow in the message's direction over weeks as the cue fades faster than the content. The sharper question it licenses is therefore not "how much did this persuade?" but "will the discounting cue and the persuasive content decay at the same rate, and which survives at retrieval?"

The decisive sharpening is that the concept locates the effect in the retrieval environment, not in the content's memory. The re-pairing test makes this crisp: re-presenting the discounting cue at the delayed measurement abolishes the effect and reverts the attitude, proving that the delayed net persuasion is the content's push minus whatever of the cue's brake remains present at recall — a source-monitoring dissociation, not a change in how the message itself was remembered. This separates three things a coarser account blurs: the persistence of what was said, the faster fading of who said it or why to doubt it, and the integration at retrieval that nets them. Holding those apart is what turns a paradoxical sign reversal into a predictable one and tells the practitioner where the lever is — to make a disclaimer stick, bind the cue to the content or re-present it at the moment of recall, because the failure is the cue's absence at retrieval, not the content being too memorable.

Manages Complexity¶

Persuasion outcomes over time form an awkward record once the immediate moment is not the only one measured: most messages decay, fading toward the baseline as memory weakens; some, delivered with a discounting cue, do the opposite, growing in the message's direction over weeks; re-presenting the cue at follow-up reverts the attitude; and adjacent source-monitoring cases show a claim surviving while its evidentiary context is lost. Treated case by case, the growing-after-rejection pattern looks paradoxical and demands its own explanation separate from ordinary decay, and the practitioner's natural reading of a low-immediate-effect campaign — "it didn't work" — closes off the measurement that would catch the reversal. The sleeper effect compresses this by replacing the single number (how much did it persuade?) with a two-trace model that generates the whole record: the persuasive content and the discounting cue are encoded as separate memory traces that decay at different rates, and the net persuasion at any time is the content's push minus whatever of the cue's brake survives. With that model, the analyst stops cataloguing decay and reversal as unrelated outcomes and reads them both off the relation between two decay curves. Cue and content fade together: ordinary decay. Cue fades faster than content: the brake lifts and net persuasion grows, inverting the sign relative to the immediate measurement. A paradox and a regularity collapse into one differential-decay account.

What the practitioner then tracks shrinks to a compact, predictive set rather than the full memory dynamics of each campaign. First, whether the message arrived with a discounting cue at all — no cue, no brake to differentially fade, so the standard decay applies. Second, the relative decay rates of the two traces: does the cue (source attribution, caveat, counter-argument) fade faster than the content, and by how much, which fixes whether and when the sign reverses. Third — the load-bearing locus — the retrieval environment: is the cue present at the moment of recall or absent, since the re-pairing test shows the effect lives in the cue's absence at retrieval, not in any change to how the content was remembered. Re-present the cue at the delayed measurement and the attitude reverts; this is what proves the integration happens at retrieval and tells the analyst exactly which variable to manipulate. From these the qualitative trajectory follows: given a cue, a faster-fading cue trace, and a retrieval environment that lacks the cue, delayed persuasion grows past the immediate figure; supply the cue at recall, or bind it to the content so it cannot fade independently, and it does not — without modelling the specific message, audience, or attitude. The compression converts the analyst's question from "how much did this persuade?" to "which of these two traces survives at retrieval?", and it hands the practitioner the matching lever directly: to make a disclaimer stick, do not make the content less memorable — make the cue persist or re-present it at recall. A confusing mix of decaying and growing persuasion outcomes becomes one two-curve rule with the retrieval moment as the decisive, manipulable parameter.

Abstract Reasoning¶

The sleeper effect licenses a two-decay-curve reading that the persuasion analyst runs in place of treating persuasion as a single number sampled at one moment. The predictive move reasons from the relation between two separately-encoded traces — the persuasive content and the discounting cue — to the net attitude at any future time: net persuasion is the content's push minus whatever of the cue's brake survives, so the analyst forecasts the trajectory by comparing decay rates rather than measuring a lump effect. The signature, counterintuitive prediction the effect underwrites is a sign reversal: where a message arrived with a discounting cue and the cue fades faster than the content, the analyst predicts attitude change will grow in the message's direction over days to weeks — the opposite of the usual decay — and predicts that a campaign reading as inert at exposure can become more persuasive later. This corrects the lazy diagnostic the field invites: a low immediate effect read as "it didn't work" stops the measurement that would catch the reversal, so the analyst reasons that a discounted message must be re-measured at delay precisely because its immediate figure can be backwards.

The locus-confirming move is interventionist and is what makes the inference clean: the analyst predicts that re-presenting the discounting cue at the moment of the delayed measurement should abolish the effect and revert the attitude — the re-pairing test — and reasons from its success that the delayed net persuasion lives in the cue's absence from the retrieval environment, not in any change to how the content was remembered. This is a source-monitoring dissociation localized by perturbation: supply the cue at recall and the brake returns; withhold it and the brake is gone. The interventionist corollary follows directly and inverts the naive lever: to make a disclaimer or warning stick, the analyst reasons that the move is not to make the content less memorable but to make the cue persist — bind it to the content so it cannot fade independently, or re-present it at the moment of recall — because the failure mode is the cue's faster decay and its absence at retrieval. The boundary-drawing move scopes the effect to its preconditions and tells the analyst when to expect it and when not: it requires (i) a discounting cue delivered with the message — no cue, no brake to differentially fade, so ordinary decay applies and the analyst predicts no sleeper pattern; (ii) the cue trace decaying faster than the content trace, with the magnitude of that gap fixing whether and when the sign reverses; and (iii) a retrieval environment lacking the cue. The compact set the analyst tracks is therefore just these three — presence of a cue, relative decay rates, and the cue's presence-or-absence at recall — and from them the qualitative trajectory (decay, or growth past the immediate figure) and the matching lever both follow, without modeling the particular message, audience, or attitude.

Knowledge Transfer¶

Within persuasion and memory psychology the effect transfers as mechanism, across the fields that rest on the same human source-versus-content memory architecture. The two-decay-curve model, the sign-reversal prediction, the retrieval-environment locus, the re-pairing test, and the inverted lever (make the cue persist, not the content less memorable) all carry intact. In persuasion research it is the Hovland wartime-film studies and their replications. In advertising it is the long-arc fade of disclaimers and credibility warnings, so consumers end up persuaded by claims they rejected at exposure. In health messaging it is a risk warning fading while the appealing content persists. In eyewitness and misinformation work it is the near-relative source-monitoring failure — a claim surviving while its discounting evidentiary context is lost — which the entry treats as the same memory architecture in a different task. Across these the diagnostics (re-measure a discounted message at delay; compare the two decay curves) and the lever move without translation, because the substrate — separable source and content traces decaying at different rates and integrating at retrieval — is the same.

Beyond message-memory the honest reading is shared abstract mechanism, not the named concept, with the important caveat that the would-be parent is emergent rather than established. The general pattern the effect instantiates is differential_decay (or cue-content dissociation): two coupled traces — a signal and its annotation, a claim and its source, a payment and its disclaimer — fade at different rates, so their net effect at time T inverts depending on which has decayed more. That pattern could in principle recur in genuinely distinct substrates — a citation network where a retraction fades but the cited claim persists, a warning label that wears off while the product appeal remains — and it is that differential-decay pattern, not "the sleeper effect," that any cross-domain lesson should carry. But the entry is candid that these echoes are sparse and each requires substantial reinterpretation, and that no such cross-domain prime is yet in the catalog; until cross-domain instances accumulate, the abstraction "outside it... collapses into a generic 'differential decay' point." What stays home-bound is everything specific to the concept: the source-versus-content memory traces, the discounting-cue mechanism, the retrieval-environment locus that the re-pairing test isolates, and the persuasion-attitude evidence base. So invoking "the sleeper effect" for a non-memory system that shows delayed sign-reversal — a retraction-without-recall in a literature, a fading caveat in a contract — is, as the entry puts it, "mostly metaphorical": it borrows the grows-after-rejection shape while dropping the human memory architecture that produces it, and the precise move is to name the generic differential_decay structure instead. The discipline is to carry the differential-decay / cue-content-dissociation pattern (acknowledging it as an emergent parent) wherever two coupled traces fade at different rates, and to reserve "sleeper effect" for the discounted persuasive message whose delayed recovery of force it actually names (see Structural Core vs. Domain Accent).

Examples¶

Canonical¶

Hovland and Weiss's 1951 experiment is the defining demonstration. Subjects read persuasive articles on topics like the feasibility of atomic submarines, with each article attributed either to a high-credibility source (e.g., a respected physicist) or a low-credibility one (e.g., a widely distrusted outlet). Measured immediately, the high-credibility source produced substantially more opinion change than the low-credibility one — the expected credibility gap. But when the same subjects were re-surveyed about four weeks later, the gap had closed: opinion change toward the low-credibility source's position had risen over time, even as the high-credibility source's advantage faded. Kelman and Hovland (1953) then supplied the locus-confirming test: reinstating the source label at the delayed measurement reinstated the discounting and abolished the sleeper gain.

Mapped back: The article carried by a distrusted attribution is the discounted message; the argument itself is the content trace and the source label is the cue trace, stored separately. The label fading faster than the argument is the differential decay; the delayed rise in agreement is the sign reversal, net persuasion growing as the brake lifts. Kelman and Hovland's reinstatement is the re-pairing test, and its success proves the retrieval-environment locus — the effect lived in the cue's absence at recall, not in a strengthened argument.

Applied / In Practice¶

The finding has shaped how persuasion researchers assess negative political advertising. Pratkanis and colleagues, whose meta-analytic work established the strict conditions under which the effect is real, connected it to attack ads: a smear paired with a discounting cue — a "paid for by [an opposing PAC]" tag, a rebuttal, or the audience's own initial skepticism — can produce little immediate attitude shift yet gain persuasive force over the following weeks as voters forget who made the claim or why they doubted it while retaining the damaging content. The practical implication inverts the naive countermeasure: a campaign cannot neutralize such an ad merely by ensuring the attack looks poorly sourced at airing; it must keep the discrediting source bound to the claim in voters' memory or re-present it near the moment of decision.

Mapped back: The attack ad tagged with a distrusted sponsor is the discounted message; the accusation is the content trace and the sponsorship/rebuttal is the cue trace. The sponsor being forgotten faster than the accusation is the differential decay, yielding delayed growth in persuasion — the sign reversal. The prescriptive lesson (re-present or bind the discrediting source near the decision) is the concept's inverted lever, targeting the retrieval-environment locus rather than trying to make the accusation less memorable.

Structural Tensions¶

T1: A striking prediction versus a fragile phenomenon (bold sign-reversal, narrow conditions). The sleeper effect's whole appeal is its counterintuitive prediction — a discounted message can grow more persuasive over weeks, inverting the usual decay — which overturns the lazy "it didn't work" reading of a low immediate effect. But the phenomenon is historically fragile: the meta-analytic record (Pratkanis and colleagues) established that it appears only under a strict conjunction — a genuine discounting cue, a cue trace that decays faster than the content, and a retrieval environment lacking the cue — and fails to replicate whenever any condition slips. The tension is that a prediction dramatic enough to warrant re-measuring every discounted campaign is also rare enough that most such re-measurements will find nothing. Lean on the effect and you chase reversals that seldom come; ignore it and you occasionally ship the attack ad that gains force. Diagnostic: Are all three preconditions actually met here — a real brake, a faster-fading cue, an absent cue at recall — or is this a discounted message that will simply decay like most?

T2: Clean retrieval-locus versus messier memory (does only the cue move). The re-pairing test localizes the effect beautifully: re-present the cue at recall and the attitude reverts, proving the delayed gain lives in the cue's absence from the retrieval environment, not in any change to the content trace. That clean dissociation is what makes the two-decay-curve model predictive. But it rests on treating the content trace as inert — merely persisting while the cue fades — when memory is reconstructive: retrieval can reshape, strengthen, or re-interpret content, and repeated recall of the argument without its brake may not just preserve but consolidate it. The tension is that the model's tractability comes from freezing one trace and letting only the other decay, whereas both traces may be changing. Trust the clean locus and the lever is sharp; admit reconstructive memory and the "content unchanged" premise is an idealization the re-pairing test does not fully rule out. Diagnostic: Does the evidence show the content trace genuinely unchanged (pure cue decay), or could repeated cue-free retrieval be strengthening or reshaping the content itself?

T3: The inverted lever versus its own backfire (re-presenting the brake reactivates the message). The concept's signature prescription inverts the naive countermeasure: to neutralize a discounted message, do not try to make the content less memorable — make the cue persist, or re-present it near the moment of decision. But re-presenting a discounting cue necessarily re-presents the claim it discounts, giving the persuasive content another exposure and another chance to be encoded, so the very act of reinstating the brake can refresh the accusation it was meant to defuse. The lever cuts both ways: bind the discrediting source to the claim and you keep the brake alive, but you also keep rehearsing the smear. The countermeasure and the harm ride on the same re-exposure, and there is no reinstating the cue without reactivating the content. Diagnostic: Would re-presenting the discrediting cue near the decision suppress the message more than the fresh exposure to the accusation strengthens it — or does reinstating the brake do net harm by rehearsing the claim?

T4: Separable traces versus integral encoding (whether the cue can fade independently at all). The mechanism presupposes that content and cue are encoded as separate traces that can decay at different rates — that is what lets the brake fade while the message persists. But encoding is not always so obliging: when a claim and its source are bound at encoding (an integral impression of "that distrusted outlet said X"), the cue does not have an independent trace to fade, and no sleeper pattern can arise. The tension is that the effect's precondition — separable traces — is itself variable and partly a matter of how the message was originally processed, so the same discounting cue produces a sleeper effect when encoded separately and ordinary decay when encoded integrally. The two-curve model assumes the separability that the phenomenon actually requires and does not always obtain. Diagnostic: Was the discounting cue encoded as a separate trace that can fade on its own, or bound integrally to the content so that it has no independent decay curve to lift the brake?

T5: Immediate figure versus the open measurement horizon (when is persuasion actually settled). The effect's deepest lesson is that a single measurement at exposure can be backwards, so a discounted message must be re-measured at delay. But this dissolves the convenient stopping point: if the immediate figure can invert, so can later ones, and the two decay curves cross at a time that depends on their unknown relative rates, so there is no principled moment at which persuasion is "finally" measured. The tension is that the concept both destroys confidence in the immediate reading and offers no closed horizon to replace it — re-measure at four weeks and the curves may still be moving, or may have crossed back. The practitioner gains the insight that timing matters and loses the assurance that any one follow-up is definitive. Diagnostic: Is there a principled decay-rate estimate that fixes when the two curves cross and stabilize, or is the chosen follow-up time an arbitrary sample of a trajectory still in motion?

T6: Autonomy versus reduction (a named persuasion effect or an emergent differential-decay pattern). "Sleeper effect" names a specific, evidence-backed phenomenon — the source-versus-content traces, the discounting-cue mechanism, the retrieval-environment locus isolated by the re-pairing test, the persuasion-attitude evidence base — and in situ, diagnosing an attack ad or a fading disclaimer, that specificity is exactly the tool. What would travel is the thinner pattern it instantiates: differential_decay / cue-content dissociation — two coupled traces fade at different rates, so their net effect inverts depending on which decayed more — which could recur in a citation network where a retraction fades faster than the cited claim, or a warning label that wears off while product appeal remains. But this is the rare case where the parent is emergent, not established: the entry is candid that cross-domain echoes are sparse and each needs heavy reinterpretation, and no such prime is yet in the catalog. The tension is that the portable structure exists in principle but is under-populated, so the reduction points at a parent still being assembled. Diagnostic: Resolve toward the (emergent) differential_decay / cue-content-dissociation pattern when two coupled traces fade at different rates in a non-memory substrate; toward the named sleeper effect when a discounted persuasive message recovers force through human source-versus-content memory in situ.

Structural–Framed Character¶

The sleeper effect sits toward the structural end but stops short of the pole — best read as mixed-structural, closely parallel to how the Simon effect is characterized: a genuine, evaluatively neutral regularity of the human source-versus-content memory system, wearing persuasion-and-memory vocabulary, and with the extra wrinkle that even its parent is only emergent. Four criteria read structural. Its evaluative_weight is nil — a discounted message recovering force over weeks is neither good nor bad; the sign reversal is a fact about how two memory traces decay, not a verdict. Its institutional_origin is none: Hovland and colleagues discovered the effect in wartime research, they did not legislate it — the differential decay of cue and content traces is a property of human memory, not an artifact of a survey or agency. And it is not human_practice_bound in the constitutive sense: it needs no institution to run, only a human perceiver whose source and content memories decay at different rates — like the Simon effect, it dissolves when its natural substrate (human message-memory) is removed, but it is a fact about that substrate rather than a practice imposed on it, and it fires whether or not anyone measures it. Within its proper substrate, cross-field reuse is recognition, not import: the same two-trace architecture is recognized intact from wartime films to advertising disclaimers to health warnings to the eyewitness-misinformation near-relative.

What keeps it off the structural pole is vocab_travels, which it fails — and unusually hard, because even the general pattern it points at is under-populated. The operative vocabulary — content trace, cue trace, discounting cue, retrieval-environment locus, the re-pairing test — is irreducibly tied to human message-memory, and beyond that substrate the entry is candid that invoking "the sleeper effect" is "mostly metaphorical," borrowing the grows-after-rejection shape while dropping the memory architecture that produces it. The portable structural skeleton is two coupled traces — a signal and its annotation, a claim and its source — fade at different rates, so their net effect at delay inverts depending on which decayed more — differential decay / cue-content dissociation. That skeleton is what the sleeper effect instantiates from its umbrella, but the umbrella is an emergent differential_decay pattern not yet in the catalog, with cross-domain echoes (a retraction fading while the cited claim persists, a warning label wearing off while product appeal remains) that are sparse and each need heavy reinterpretation. So such cross-domain reach as exists belongs to that emergent parent, while the source-versus-content specialization, the discounting-cue mechanism, and the persuasion-attitude evidence base that make "the sleeper effect" the specific named effect stay pinned to human memory. Its character: structural in skeleton — a real, evaluatively neutral, recognized-in-nature memory mechanism — but stated in message-memory vocabulary that pins it to human cognition, leaving it mixed-structural, the persuasion instance of an as-yet-emergent differential-decay pattern rather than a free-floating prime.

Structural Core vs. Domain Accent¶

This section decides why the sleeper effect is a domain-specific abstraction and not a prime, and it carries the case for its domain-specificity — there is no separate section for that. It is an unusually clean case, with one extra wrinkle: even the parent it would reduce to is only emergent.

What is skeletal (could lift toward a cross-domain prime). Strip the persuasion-memory content and a thin relational structure survives: two coupled traces — a signal and its annotation, a claim and its source, a payment and its disclaimer — fade at different rates, so their net effect at a delayed reading inverts depending on which trace decayed more. The pieces that travel are abstract — a persistent component, an annotating component that brakes or discounts it, differential decay of the two, and a net that is read as the persistent push minus whatever of the brake survives, with a sign reversal whenever the brake fades sufficiently faster. That skeleton is genuinely portable in principle and could recur in distinct substrates: a citation network where a retraction fades faster than the cited claim, a warning label that wears off while the product appeal remains. It is the core the sleeper effect shares, not what makes it distinctive — but see the wrinkle below.

What is domain-bound. Almost everything that makes the concept the sleeper effect in particular is human message-memory furniture that does not survive extraction: the source-versus-content memory traces encoded separately; the discounting-cue mechanism (low-credibility source, counter-argument, disclaimer); the retrieval-environment locus — the effect living in the cue's absence at recall rather than in any change to the content trace; the re-pairing test that isolates that locus by reinstating the cue at delayed measurement; and the persuasion-attitude evidence base (Hovland's wartime films, the Kelman–Hovland reinstatement, the negative-advertising meta-analyses). The decisive test: remove human message-memory and there is no separable content trace and cue trace to decay at different rates — no brake that can fade independently of what it braked — so "the sleeper effect" has nothing to name; off-substrate, as the entry says, invoking it is "mostly metaphorical."

Why this does not clear the prime bar. A prime is a relational structure whose vocabulary travels and whose cross-domain transfer is recognition of the same mechanism, not analogy. The sleeper effect's transfer is bimodal. Within persuasion and memory psychology it travels intact as mechanism — the two-decay-curve model, the sign-reversal prediction, the retrieval-environment locus, the re-pairing test, and the inverted lever (make the cue persist, not the content less memorable) all carry from wartime-film studies to advertising disclaimers to health warnings to the eyewitness-misinformation near-relative, because the substrate (separable source and content traces integrating at retrieval) is shared. Beyond that substrate it travels only by analogy: invoking "the sleeper effect" for a non-memory system that shows delayed sign-reversal — a retraction-without-recall in a literature, a fading caveat in a contract — borrows the grows-after-rejection shape while dropping the human memory architecture that produces it. And when the bare structural lesson is needed cross-domain, the honest move is to carry the general differential_decay / cue-content-dissociation pattern instead — with the caveat that this parent is emergent, not established: the entry is candid that cross-domain echoes are sparse and each requires heavy reinterpretation, and no such prime is yet in the catalog. So the sleeper effect stays below the prime bar on two counts at once: its distinctive vocabulary is pinned to human message-memory, and even the portable structure it points at is a parent still being assembled rather than one that could yet carry the cross-domain reach on its own.

Relationships to Other Abstractions¶

Current abstraction Sleeper Effect Domain-specific

Parents (1) — more general patterns this builds on

Sleeper Effect is a kind of Differential Decay Prime

The Sleeper Effect is the persuasion-memory species of Differential Decay, where a discounting cue fades faster than message content and delayed net persuasion rises or reverses.

Hierarchy path (1) — routes to 1 parentless root

Sleeper Effect → Differential Decay → Signal Decay and Fadeout → Dissipation → Irreversibility → Reversibility and Irreversibility

Not to Be Confused With¶

Ordinary persuasion decay. The usual time course in which attitude change fades toward baseline as memory weakens. The sleeper effect is the opposite — a sign reversal in which delayed persuasion grows past the immediate figure. Tell: does the attitude change shrink over weeks (ordinary decay) or rise in the message's direction (sleeper effect)? The two differ in the sign of the time trend, and the sleeper case requires a discounting cue to have been present.
Mere exposure effect. The rise in liking produced by repeated contact with a stimulus, independent of content. The sleeper effect is about discounted content recovering persuasive force as the discount fades — a cue-content dissociation over time, not a familiarity boost from repetition. Tell: is the driver repeated exposure raising bare liking (mere exposure) or a single discounted message whose brake decays faster than its content (sleeper)? Mere exposure needs no discounting cue; the sleeper effect requires one.
Illusory truth effect. The finding that repeated statements are judged more true, because fluency is misread as accuracy. It resembles the sleeper effect in yielding delayed belief in dubious claims, but its engine is repetition-driven fluency, not the differential decay of a separately-stored source cue. Tell: is the claim believed because it was heard many times (illusory truth) or because its discounting cue was forgotten faster than the claim itself (sleeper)? One needs repetition; the other needs a fading brake.
Misinformation effect. The near-relative in eyewitness memory: post-event information contaminates memory for an event. It shares the sleeper effect's source-versus-content architecture (content surviving while its discounting context is lost) but concerns memory for events, not attitude change from a persuasive message. Tell: is the task remembering what happened, corrupted by later input (misinformation), or being persuaded by a claim whose source-discount faded (sleeper)? Same memory machinery, different task and outcome measure.
Source amnesia / source-monitoring failure (the general phenomenon). The broad failure to recall where a memory came from while retaining its content. The sleeper effect is the persuasion-specific case where that source-loss lifts a discounting brake and increases attitude change. Tell: are you naming the generic inability to attribute a memory to its source (source amnesia) or the specific delayed persuasion gain it produces when the lost source was a discounting cue (sleeper effect)? The former is the mechanism; the latter one consequence of it.
The emergent differential_decay / cue-content-dissociation pattern (the would-be parent). The substrate-neutral structure — two coupled traces fade at different rates, so their net effect inverts depending on which decayed more — that the sleeper effect instantiates in human memory. It is what any genuine cross-domain lesson should carry (a retraction fading while a cited claim persists), but the entry is candid it is emergent, not yet catalogued. Tell: strip the human source-versus-content memory and what remains — two coupled decaying quantities whose net inverts — is this general pattern, not "the sleeper effect," whose cross-domain use is otherwise mostly metaphorical. (Treated fully in a later section.)

Neighborhood in Abstraction Space¶

Sleeper Effect sits in a crowded region of the domain-specific corpus (33^rd percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Persuasion & Rhetorical Influence (13 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-07-12