Social Priming

What it is

Two quite different things travel under the word priming, and almost all the confusion around this topic comes from letting them blur. The first is a measured, undramatic effect on processing. Show someone the word "nurse" and they will recognise the word "doctor" a few milliseconds faster than they otherwise would. That is semantic priming. Repetition priming works the same way: having seen a thing recently makes it faster to see again. These effects are real, they replicate, they are measured in reaction time, and nobody disputes them. They are also modest and boring, in the way that findings which survive tend to be.

The second is a much larger claim, and it is the one this page is about. Behavioural social priming holds that an incidental cue, glimpsed and not consciously registered, changes what a person subsequently does in the world. Not what they recognise faster, but how generous they are, how attractive they find a stranger, how warmly they judge someone's character. The intervention is tiny and momentary; the claimed effect is a change in real action. That shape, big behavioural effect from a small incidental exposure with no plausible dose-response, turns out to be the reliable signature of a finding that will not survive being run again.

The collapse of this literature is unusually well documented, because it was the field's most-cited work and therefore the first thing replicators went after. Three results in particular did the reputational damage.

Williams and Bargh's 2008 finding that holding a warm cup makes people judge a stranger as personally warmer became, for a while, the single most quoted result in social psychology. Chabris and colleagues replicated it with triple the original sample, double-blind, and found effects close to zero, r = −.03 and .02. Three further high-powered replications, 861 participants between them, found nothing at all.

Money priming, Vohs and colleagues' 2006 claim that reminders of cash make people less helpful and more self-sufficient, failed across four high-powered replications. Questionable research practices were later identified in the original studies.

"Romantic red", Elliot and Niesta's 2008 finding that a red frame raises perceived attractiveness, returned d = 0.09 for men and d = −0.09 for women under pre-registered replication, effects in opposite directions and close enough to zero to be noise. A nine-study meta-analysis covering 640 participants found no effect.

The aggregate picture across the wave of large-sample replications of this literature is the clearest number the replication crisis produced: 94 per cent of replications yielded smaller effects than the original studies, and only 17 per cent reached statistical significance in the original direction. Those figures describe social priming specifically, not psychology as a whole, and the distinction is the whole point of stating them.

It is worth being careful about what this failure means. It is not, in the main, a story about fraud. It is a story about a research culture in which small samples, publication bias and enough analytic latitude to coax a null result into a publishable one combined to produce a literature of findings that were never there. The single strongest predictor of whether one of these replications succeeded was not sample size or design. It was whether an original author sat on the replication team. That points at unconscious researcher latitude rather than anything deliberate.

The reason this concept still earns a page, rather than simply being deleted, is that a great deal of popular psychology writing was built on it and is still in circulation. Adam Alter's Drunk Tank Pink draws its evidentiary spine, the "within us, between us, around us" cue-to-behaviour argument, directly from this literature. The book is a snapshot of a field taken just before the reckoning reached it. Anyone reading it, or reading the enormous body of marketing and design advice downstream of it, needs to know which tier of evidence they are standing on. See The Replication Crisis for the wider pattern, and The Context Effect for the sturdier neighbouring claim that is routinely and wrongly filed alongside it.

In effect

Research: the warm cup, and what happened when it was run properly

Williams and Bargh (2008) had participants briefly hold a hot or iced drink, then rate a stranger described in a short vignette. Those who had held the warm cup rated the stranger as warmer in personality. It is a beautiful result: an abstract metaphor apparently grounded in a physical sensation. Chabris and colleagues ran it again with triple the sample under double-blind conditions and found r = −.03 and .02, which is nothing. Three further high-powered replications totalling 861 participants also found nothing. Bargh has published his own response to the replications, which is part of the record. The honest summary is that this specific finding did not hold, and that it should not be repeated as fact in any article, deck or design rationale.

Everyday life: "nurse" then "doctor", which is real and is not this

Read a list of words containing "nurse", then be asked to say whether "doctor" is a word. You will be measurably faster than if the list had contained "butter". That is semantic priming, it happens constantly while reading, and it is not in dispute. It is also the reason people find the behavioural claims plausible: the small version of the phenomenon is genuinely true, which lends unearned credibility to the large version. Keeping the two apart is the single most useful thing a reader can do with this topic. Processing is primed. Behaviour, on the current evidence, is not.

Commerce and product design: red packaging, and the advice built on it

"Romantic red" travelled from a 2008 paper into a decade of design and marketing advice about colour, attraction and conversion. Elliot and Niesta's finding failed pre-registered replication, d = 0.09 for men and d = −0.09 for women, and a nine-study meta-analysis of 640 participants found no effect. The commercial cost of this is specific rather than abstract. Colour, packaging and cue-based conversion advice sold on the strength of that literature has no evidential floor under it, and a brand that has justified a design decision by citing it has cited a finding that is not there. The practical test to apply before repeating any such claim is the one the replication record supports: big effect, tiny incidental intervention, no dose-response, treat as unsupported until shown otherwise.

What it does not say

It does not say that priming as a whole is bogus. Semantic and repetition priming are robust and were never the problem. What collapsed is specifically the behavioural social-priming literature, the claim that incidental exposure changes action rather than reaction time. Anyone using the failure of the second to dismiss the first has made the mirror image of the original error.

It does not say that psychology as a field is unreliable. The 94 per cent and 17 per cent figures describe social-priming replications specifically. Applying them to findings outside that literature would repeat exactly the overgeneralisation this page exists to prevent. Several neighbouring effects, among them the misinformation effect and social facilitation, survived the same scrutiny intact.

Bargh asked real questions with the methods available at the time, and those methods were not strong enough to answer them. His published response to the replications is part of the record. The vault's position on him is care rather than dismissal. What follows from the record is a hard rule about citation, not a verdict on a person: nothing from this literature reaches a published article here without its replication status being checked first.

It does not, finally, settle every claim in the vicinity. This page is tagged contested, which in this publication's vocabulary means the finding has failed replication or is actively disputed in the primary literature, and that it may still be worth writing about while not being citable as fact. The originator field records John Bargh at medium confidence. He is the field's central figure and the author of its most-replicated failure, which is not quite the same as being the sole inventor of the idea, and the page should be read accordingly.

Who developed it

John Bargh, a social psychologist at Yale, is the central figure in the behavioural social-priming literature and the co-author, with L. E. Williams, of its most famous and most decisively failed result. The vault records him as originator at medium confidence, with a date of 2006, which reflects his centrality to the field rather than a claim that he invented priming research. Kathleen Vohs led the money-priming work, and Adam Alter is the writer who carried most of this literature into the popular market.

---