John Bargh

Who they are

Bargh is the James Rowland Angell Professor of Psychology at Yale, having previously spent much of his career at New York University. The question he has worked on for four decades is a real one and predates any of the controversy: how much of ordinary social behaviour runs without conscious intention. Nobody sensible disputes that some of it does. What is disputed is how far the effect reaches, and whether the experiments that appeared to demonstrate it were measuring anything.

Two results carried the argument, and both have come apart in public.

The first is elderly priming. Bargh, Chen and Burrows reported in 1996 that participants who had unscrambled sentences containing words associated with old age walked more slowly down the corridor afterwards. The effects were enormous, above one standard deviation across two studies of about thirty people each, and the finding became a fixture of textbooks, lectures and popular writing about the unconscious. In 2012 Doyen, Klein, Pichon and Cleeremans published a replication in PLoS ONE using automated timing and a larger sample. The priming words did nothing to walking speed. Their second experiment then manipulated what the experimenters themselves believed was about to happen, and the slowing appeared only among the experimenters who expected it. That is a finding about measurement rather than about old age, and it is the reason this particular failure did the damage it did.

The second is warmth. Williams and Bargh reported in Science in 2008 that briefly holding a hot drink made people rate a stranger as warmer in personality, an abstract metaphor apparently grounded in a physical sensation. Chabris and colleagues ran it again with triple the original sample under double-blind conditions and found effects of r = −.03 and .02. Three further high-powered replications, 861 participants between them, found nothing at all. Note the attribution, because it is routinely dropped: L. E. Williams is first author on the 2008 paper, and it is filed everywhere under Bargh's name alone.

Bargh has published his own response to the replications, and it is part of the record rather than a footnote to it. His argument is that priming effects are conditional on context, subtle in operation, and easily extinguished by a replication team that does not run the procedure as the original lab ran it, which makes a null result uninformative rather than decisive. Replicators answer that a finding which only appears in the hands of people who expect it is not a finding. The disagreement has not been settled by either side conceding.

Two things follow from this, and they pull in different directions.

The first is a hard rule about citation. Nothing from Bargh's literature reaches a published article on this site without its replication status being checked first. That is a standing instruction rather than a preference, and it exists because a great deal of trade psychology, Drunk Tank Pink among it, was built on this material at the moment the field's confidence was highest.

The second is that dismissal would be the wrong reading. Bargh asked real questions, and the methods available when he asked them were not strong enough to answer them. Small samples, publication bias and a research culture that rewarded surprising results produced a literature of effects that were probably never there, and that description covers most of a generation of social psychology rather than one laboratory. Our notes pair him with Alexander Schauss, who has no page on this site yet, for exactly this reason: one from the fringe journals of the 1970s, one from peer-reviewed Science in the 2000s, both overturned by later scrutiny, which says something about how little the venue protected either of them.

Books

One book, published in 2017.

It appeared five years after the Doyen replication and nine after the warmth study, so the failures described above were already in the literature when it was written. Our notes do not record a reading of it, and this page therefore makes no claim about how it handles them.

Key papers

Bargh, J. A., Chen, M., & Burrows, L. (1996). "Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action." Journal of Personality and Social Psychology, 71(2), 230-244. First author. The elderly-priming study. Participants who unscrambled sentences containing words associated with old age walked more slowly down the corridor afterwards, across two studies of roughly thirty people each, with very large reported effects.

Doyen, S., Klein, O., Pichon, C.-L., & Cleeremans, A. (2012). "Behavioral priming: It's all in the mind, but whose mind?" PLoS ONE, 7(1), e29081. The replication, with automated timing and a larger sample, in which the priming words did not change walking speed. A second experiment varied what the experimenters themselves expected, and the slowing appeared only where they expected it, which makes it a result about measurement.

Williams, L. E., & Bargh, J. A. (2008). "Experiencing physical warmth promotes interpersonal warmth." Science, 322(5901), 606-607. Williams is first author, and the paper is filed under Bargh's name almost everywhere. Briefly holding a hot drink was reported to raise ratings of a stranger's personal warmth. Chabris and colleagues ran it at triple the sample under double-blind conditions and found effects indistinguishable from zero, and three further high-powered replications totalling 861 participants found nothing.

Bargh's own published reply is part of this list rather than a footnote to it. His position is that priming effects are conditional on context, subtle to produce, and easily extinguished by a team not running the procedure as the originating laboratory ran it, so a null result is uninformative rather than decisive. The replicators' answer is that an effect appearing only in the hands of people expecting it is not yet an effect. Neither side has conceded, and this page records the disagreement as a live scientific one rather than settling it.

Key ideas

Automaticity: a substantial share of social behaviour is initiated without conscious intention, and the person acting cannot report what set it off.

Behavioural priming: an incidental cue, glimpsed and not registered, changes not what a person notices but what they do. This is the claim that failed.

Conceptual metaphors are grounded in physical experience, so warmth in the hand becomes warmth in a judgement. That was the mechanism proposed for the 2008 result.

His position on the replications: priming effects are context-dependent and fragile, so a null result from a lab running an unfamiliar procedure does not establish absence.

The counter-position, which the record supports better: the strongest predictor of a successful replication in this literature was whether an original author sat on the replication team.

Concepts on this site

Social Priming, the literature he is central to, and the worst-performing subfield of the replication crisis

The Replication Crisis, the wider pattern, where his warmth study is the worked example

The Context Effect, the sturdier neighbouring claim that is routinely and wrongly filed alongside priming

---