Social Psychology
Sleeper Effect
Why a discounted message can become more persuasive weeks after you heard it
The sleeper effect is a counterintuitive persuasion phenomenon in which a message paired with a low-credibility source — initially dismissed — gains influence over time as the source is forgotten but the content is retained. Hovland and Weiss documented it during World War II propaganda research at Yale (1951). The effect emerges roughly four to six weeks after exposure and requires a discounting cue strong enough to suppress immediate persuasion. Kumkale and Albarracín's 2004 meta-analysis of 72 studies confirmed it is real but fragile, occurring under specific conditions of cue strength, message strength, and timing.
- DocumentedHovland & Weiss (1951), Yale Communication Program
- Time courseReverses 4-6 weeks post-exposure
- Meta-analysisKumkale & Albarracín (2004), 72 studies
- Required ingredientStrong discounting cue + strong message
- MechanismSource-message dissociation in memory
- MagnitudeModest reversal (d ≈ 0.08-0.30)
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why the sleeper effect matters
- Political advertising. Attack ads dismissed today may shape opinion in a month.
- Misinformation control. Retractions decay faster than the false claims they correct.
- Health communication. Discredited fringe claims can resurface in patient decisions.
- Marketing. Skeptical first impressions soften as the brand becomes familiar.
- Legal proceedings. Inadmissible evidence sometimes lingers in juror memory.
- Education. Source-tagged content benefits from re-exposure to maintain attribution.
- Propaganda research. Foundational case study for time-delayed influence.
Common misconceptions
- Persuasion always grows with time. Most messages decay; the sleeper effect is an exception requiring specific conditions.
- Any low-credibility source produces it. The discounting cue must fully suppress immediate persuasion or the effect never emerges.
- It's a large effect. Meta-analytic effect sizes are small (d ≈ 0.08-0.30).
- Forgetting the source is enough. Differential decay — source fading faster than content — is the mechanism, not pure forgetting.
- It was disproven in the 1970s. Pratkanis et al. (1988) reframed the conditions and rescued the phenomenon.
- Cue placement does not matter. Cues processed after the message produce sleeper effects; cues before do not.
Frequently asked questions
What's the original sleeper effect study?
Hovland and Weiss (1951) had participants read articles attributed to either high-credibility sources (Robert Oppenheimer on atomic submarines) or low-credibility sources (Pravda). Immediately after, high-credibility sources were more persuasive. Four weeks later, the gap closed and in some cases reversed — the low-credibility message had retained influence while the source label faded. The Yale Communication Program ran the work as part of WWII propaganda research.
Why does the effect occur?
The leading account is the dissociation hypothesis — source and content decay at different rates in memory. The content-based attitude change is encoded as a semantic memory; the source attribution is a meta-tag that fades faster. By week four, you remember "atomic submarines might work" without remembering you read it in Pravda. Pratkanis et al. (1988) refined this into the differential decay model.
What conditions are required?
Pratkanis, Greenwald, Leippe, and Baumgardner (1988) showed three conditions must hold. (1) The message must be strong enough to produce attitude change absent the cue. (2) The discounting cue must be powerful enough to fully suppress immediate persuasion. (3) The cue must be processed after the message, not before. Without all three, no sleeper effect emerges — explaining decades of mixed replications.
Did the effect replicate?
Mixed history. Gillig and Greenwald's "Is it time to lay the sleeper effect to rest?" (1974) reported failures and nearly killed the field. Pratkanis et al. (1988) revived it with the differential decay framework. Kumkale and Albarracín's 2004 meta-analysis of 72 experiments confirmed a real but small effect (d ≈ 0.08 absolute, 0.30 under ideal conditions). Today it is considered real but narrow.
How is it used in propaganda?
Negative political ads exploit it. The classic pattern — air a damaging claim, see it dismissed when sourced to a fringe outlet, then watch it surface in opinion polling weeks later — fits sleeper logic. Tabloid stories about politicians, conspiracy claims pre-empted by fact-checks, and disclaimers attached to advertorials all create source-message dissociation that may produce delayed persuasion.
How does it relate to misinformation?
Major implication. A retracted news story can still influence beliefs weeks after the retraction, because the original claim is encoded as a memory while the retraction is encoded as a tag. Lewandowsky et al.'s continued influence research (2012) shows corrections only partially erase initial misinformation. Sleeper logic suggests labels — "this came from an unreliable source" — weaken faster than the content they label.
How can you defend against it?
The strongest defense is integrating the source into the content itself, not appending it as a postscript. Telling viewers "Pravda claims X" while reading X embeds the source in the encoded representation. Forewarning before exposure works better than disclaiming after. Spaced re-exposure to the source-content link refreshes the tag. Skilled debunkers repeat the corrected fact, not the misinformation, to avoid familiarity-driven retention.