AI tools helped gather research and write the first draft of this article. I edited it and checked every study figure against the paper's official abstract, listed at the end. The cover illustration is AI-generated. Details at the end.
In 1975 Robert Cialdini and his colleagues published three experiments on a technique they called the door-in-the-face. Ask for an extreme favour, let the person refuse, then ask for a smaller one. In all three experiments, that sequence produced more agreement than asking for the smaller favour alone [1].
Forty-five years later, a team led by Oliver Genschow repeated the study as closely as they could, in another country on another continent. Compliance rates were "similarly high" to the original [2]. The authors concluded that at least some findings in social psychology can transcend a particular time, place and population [2].
That sounds like a law. But in 2012 Thomas Feeley, Ashley Anker and Ariel Aloe combined 35 years of door-in-the-face studies. The technique had a clear effect on verbal compliance, people saying yes (78 studies, r = .126). On behavioural compliance, people actually doing the thing, the effect was smaller and not statistically significant (39 studies, r = .052) [3].
So a tactic that survived a direct replication after 45 years reliably changes what people agree to, and much less reliably what they do. That is the difference between a tendency and a law, and it is what this article is about.
This is the fifth part of Influence & Incentives, a series on how other people, rules and rewards shape what we decide, write and publish. The earlier parts looked at social proof, reciprocity, scarcity and rewards one at a time. This one steps back and asks how much weight any of these "principles" can carry.
What a principle is, and what it is not
Robert Cialdini's book Influence groups persuasion research into a short list of principles: reciprocity, commitment and consistency, social proof, liking, authority, scarcity and, in the 2021 edition, unity. The list is memorable, and that is part of its success. It is also easy to misread.
A law of physics holds everywhere, with a known size, under stated conditions. A persuasion principle is a name for a family of findings. Each finding has three properties that a slogan hides:
- A size. Some effects are large, many are small.
- An outcome. Some studies measured what people did; many measured what they said or intended.
- Conditions. Topic, timing, the request, the audience and how hard the request is to grant all change the result.
The evidence behind the principles is also uneven. Some findings have been repeated by independent teams decades later. Others have shrunk or disappeared when large groups of laboratories tried them again. The sections below sort a few well-known examples into three groups: held up, smaller than their reputation, and did not replicate.
Effects that held up
The door-in-the-face, at least for agreement. The 2021 direct replication found that people who turned down an extreme request were more likely to agree to the target request than people who heard only the target request, or a similarly small request first [2]. That is a strong result for a 45-year-old finding.
Obedience to authority, in a limited form. In 2009 Jerry Burger published a partial replication of Stanley Milgram's obedience studies [6]. For ethical reasons, his 70 adult participants were stopped at 150 volts, the point where the "learner" first protests. Burger chose that point because 79% of Milgram's participants who went past it continued to the end of the generator's range [6]. His obedience rates were "only slightly lower" than Milgram's 45 years earlier [6].
The low-ball. Here a person agrees to a request, and then the cost is raised slightly. A 2015 meta-analysis by Burger and Deanna Caputo found it to be a reliable way to increase compliance, most effective when people state their first agreement publicly and when the second request is only slightly more costly [5].
Even these come with conditions. The low-ball works best under specific circumstances [5]. Burger's study stopped at the first protest, so the full obedience rate was estimated, not observed [6]. And the door-in-the-face is strongest on words [3].
Effects smaller than their reputation
In 1984 James Dillard, John Hunter and Michael Burgoon combined the studies on two sequential-request tactics: the door-in-the-face and its mirror image, the foot-in-the-door, where a small first request is meant to make a larger second one easier [4]. Their summary: both effects were small (r = .17 for foot-in-the-door, r = .15 for door-in-the-face), "even under optimal conditions" [4].
The conditions mattered as much as the size. Both tactics needed a prosocial topic to work. The door-in-the-face worked only when the second request came soon after the first [4].
The 2012 meta-analysis added more conditions [3]. The door-in-the-face worked better for requests to volunteer or take part in research than for other requests, such as giving money. And in settings where people rarely agreed anyway, which the authors describe as tougher contexts (lower baseline compliance), the effect was smaller [3].
Put together, the honest description of the door-in-the-face is not "refusal creates obligation". It is closer to this: a short gap between requests, a good cause and a modest second request can raise the share of people who say yes, by a small amount, and less reliably the share who follow through.
Effects that did not replicate
The foot-in-the-door, in one direct test. Malgorzata Gamian-Wilk and Dariusz Doliński repeated the original 1966 foot-in-the-door experiment by Jonathan Freedman and Scott Fraser in Poland and in Ukraine, in 2003 and again in 2013 [7]. In 2003 the tactic worked in Ukraine and did not work in Poland. In 2013 the effect was not significant in either country. The authors also note that the effect size of the original studies "has never been obtained again" [7]. One replication does not settle a question, but it shows that a famous result can depend on place and time.
Behavioural priming. In a widely cited 1996 study by John Bargh and colleagues, participants unwittingly exposed to the stereotype of old age walked more slowly when leaving the laboratory, as Stéphane Doyen and his co-authors summarise it [8]. In 2012 Doyen's team tried to repeat it with automated timing and a larger sample. Their first experiment showed no priming. In a second experiment they told some experimenters that primed participants would walk slower and others the opposite. The slower walking appeared only when the experimenters expected it [8].
Priming knowledge. A 1998 study reported that people primed with the idea of a "professor" scored 13% better on a trivia test than people primed with "soccer hooligans". In a Registered Replication Report published in 2018, 23 laboratories that met all the inclusion criteria tested 4,493 people. The difference between the conditions was 0.14%, and the authors observed no overall difference [9].
Being mimicked. Earlier studies suggested that people who are subtly mimicked feel closer to the person mimicking them and behave more generously. In 2023 Birgit Rauchbauer and colleagues re-examined this with 180 female participants. The participants' own empathy-related traits predicted closeness and helping better than being mimicked did [10].
Priming matters for this series because the same idea is often sold to marketers, sometimes under Cialdini's later label "pre-suasion": a subtle cue placed before a request changes the answer. The two direct tests above did not support the dramatic versions of it.
A pattern from many laboratories at once
Single replications can be argued with. Large multi-lab projects are harder to dismiss, and they say something unexpected about context.
In the first Many Labs project, published in 2014, 36 samples with 6,344 participants tested 13 effects. Ten replicated consistently, one weakly, and two priming effects did not replicate. Whether a sample was tested in a lab or online, or in the US or elsewhere, by and large did not predict the size of the effect [11].
Many Labs 2, published in 2018, went further: 28 published findings, 125 samples, 15,305 participants from 36 countries and territories [12]. Depending on the threshold, 14 or 15 of the 28 replicated, about half. Three in four replications produced smaller effects than the originals, and the median effect size fell from 0.60 to 0.15 [12]. Differences between samples, including comparisons between more and less "WEIRD" cultures (Western, educated, industrialised, rich, democratic), explained little. The authors attributed the variation more to the effect being studied than to the sample or setting [12].
This matters for anyone who hears "that study does not apply here, our audience is different". Sometimes that is true; the Polish and Ukrainian foot-in-the-door results suggest place and time can matter [7]. But the multi-lab evidence points to a more common problem: many original effects were smaller, or weaker, than first reported, wherever they were tested.
Why this matters for crypto and growth marketing
Persuasion principles travel well as content. Threads and courses present them as "psychological triggers" with guaranteed results, and crypto marketing borrows the same language for launches, communities and token sales.
The research above gives three reasons to discount those promises.
First, many classic studies measured agreement or intention, while a purchase, a wallet connection or a token swap is behaviour. For the door-in-the-face, the effect on behaviour was less than half the effect on agreement and not statistically significant [3].
Second, the techniques most often sold as clever, subtle cues and mimicry, had the weakest record among the studies above [8][9][10]. The ones that held up are blunt: direct requests, visible commitments, clear authority.
Third, even reliable effects were small on average [4][12]. A small effect can still matter across millions of impressions. It is rarely large enough to rescue a weak product or to justify a promise of "10x".
Four questions that separate a tendency from a law
When someone cites a persuasion principle, these four questions recover what the slogan left out:
- What was measured: words or actions? Saying yes, rating a product or stating an intention is not the same as paying, signing up or showing up [3].
- How big was the effect, and compared with what? "Significant" means unlikely to be zero, not large. Ask for the size and the control condition [4][12].
- Has an independent team repeated it, ideally many teams with a plan registered in advance? The professor-priming result did not survive 23 laboratories [9]; the door-in-the-face did survive a direct replication [2].
- How close are the conditions to yours? Topic, timing, how hard the request is and who is asked all changed the results [3][4].
Five checks before you trust a persuasion "hack"
- Is there a named study? "Psychology says" is not a source. A paper title, an author and a year are the minimum.
- Was it behaviour? If the study counted clicks, purchases or real sign-ups, it is closer to your problem than a questionnaire.
- Was it repeated? Search the study's name with "replication". Many famous effects now have one.
- Does the tactic depend on people not noticing it? Subtle cues were the weakest group in the evidence above, and tactics that rely on hiding the mechanism also raise an honesty question.
- What would a fair test look like for you? If the answer is "we never tested it, it is a known law", it has not been tested.
A worked example (fictional)
The project, the consultant and all figures below are invented to show the checks in use.
A growth consultant pitches a new token community: "Use the foot-in-the-door. First ask people to like a post, then to join the channel, then to buy in the presale. Psychology proves each small yes makes the next one automatic."
Run the questions. The principle is real but the claim attached to it is not: the meta-analysis found the effect small even under optimal conditions, and only for prosocial topics [4]. A direct replication found no significant effect in 2013 [7]. A presale is not a prosocial request, and buying is behaviour, not agreement. Nothing in the pitch names a study, an effect size or a test.
A version worth considering would say: "We think a small first step may raise sign-ups. We will test it: half of new visitors see the step, half do not, for four weeks, and we will report sign-ups, not likes." That is a hypothesis with a measurement attached, which is what a persuasion principle should be.
Using the principles honestly
For anyone who writes or publishes, including this site, the research suggests a few habits:
- Treat a principle as a reason to test, not as a result. Measure the behaviour you care about.
- Report the size of an effect, not only its direction, and say what it was compared with.
- Prefer the tactics that held up because they are open: a clear request, a real source, a commitment people choose to make in public.
- Do not build on subtle cues that the evidence does not support, and do not use any tactic that only works if the reader does not know about it.
What this article cannot tell you
The studies here are a selection, chosen because they are well known or because they were repeated at scale. They do not cover every persuasion technique, and a failed replication of one version of an effect does not prove that no version works. Most of the research was done in laboratories, with students or online panels, on requests that are not about money; none of it measures crypto audiences or readers of Dapp.Expert. I relied on official abstracts, not full papers, so details such as exact compliance percentages per condition are left out where the abstracts do not give them. The effect sizes quoted are averages across very different studies and should not be turned into a forecast for any single campaign.
How this article was made
| Task | What the AI did | What a person did | How it was checked |
|---|---|---|---|
| Research | Collected studies on influence and online behaviour into research notes | Chose the question and the studies to rely on | Figures and quotations checked against the official abstracts of the papers [1]–[12] |
| Draft | Wrote the first draft from the notes | Edited the text, the structure and the worked example | A separate review compared figures and claims with the sources |
| Images | Generated the cover photograph-style illustration | Chose the illustration; designed the explanatory diagrams | Labelled as AI-generated in the disclosure line and alt text |
| Not done by AI | — | The decision to publish and responsibility for the text | — |
Conflicts of interest: none. The author has no commercial relationship with the researchers or organisations named; the project in the example is fictional.
Responsible for this article: Dmitriy Lastovka, author. Corrections: contact form.
Sources
Labels: paper (peer-reviewed study, read via its official abstract).
- Robert B. Cialdini, Joyce E. Vincent, Stephen K. Lewis, José Catalan, Diane Wheeler, Betty Lee Darby, "Reciprocal concessions procedure for inducing compliance: The door-in-the-face technique", Journal of Personality and Social Psychology 31(2) (1975), 206–215. https://doi.org/10.1037/h0076284 Paper, official abstract.
- Oliver Genschow et al., "Does social psychology persist over half a century? A direct replication of Cialdini et al.'s (1975) classic door-in-the-face technique", Journal of Personality and Social Psychology 120(2) (2021), e1–e7. https://doi.org/10.1037/pspa0000261 Paper, official abstract.
- Thomas Hugh Feeley, Ashley E. Anker, Ariel M. Aloe, "The Door-in-the-Face Persuasive Message Strategy: A Meta-Analysis of the First 35 Years", Communication Monographs 79(3) (2012), 316–343. https://doi.org/10.1080/03637751.2012.697631 Paper, official abstract.
- James Price Dillard, John E. Hunter, Michael Burgoon, "Sequential-request persuasive strategies: Meta-analysis of foot-in-the-door and door-in-the-face", Human Communication Research 10(4) (1984), 461–488. https://doi.org/10.1111/j.1468-2958.1984.tb00028.x Paper, official abstract.
- Jerry M. Burger, Deanna D. Caputo, "The low-ball compliance procedure: a meta-analysis", Social Influence 10(4) (2015), 214–220. https://doi.org/10.1080/15534510.2015.1049203 Paper, official abstract.
- Jerry M. Burger, "Replicating Milgram: Would people still obey today?", American Psychologist 64(1) (2009), 1–11. https://doi.org/10.1037/a0010932 Paper, official abstract.
- Malgorzata Gamian-Wilk, Dariusz Doliński, "The Foot-in-the-Door Phenomenon 40 and 50 Years Later: A Direct Replication of the Original Freedman and Fraser Study in Poland and in Ukraine", Psychological Reports 123(6) (2020; online 2019), 2582–2596. https://doi.org/10.1177/0033294119872208 Paper, official abstract.
- Stéphane Doyen, Olivier Klein, Cora-Lise Pichon, Axel Cleeremans, "Behavioral Priming: It's All in the Mind, but Whose Mind?", PLOS ONE 7(1) (2012), e29081. https://doi.org/10.1371/journal.pone.0029081 Paper, official abstract.
- Michael O'Donnell et al., "Registered Replication Report: Dijksterhuis and van Knippenberg (1998)", Perspectives on Psychological Science 13(2) (2018), 268–294. https://doi.org/10.1177/1745691618755704 Paper, official abstract.
- Birgit Rauchbauer, Gabriela Jank, Robin Dunbar, Claus Lamm, "Only empathy-related traits, not being mimicked or endorphin release, influence social closeness and prosocial behavior", Scientific Reports 13 (2023), 4072. https://doi.org/10.1038/s41598-023-30946-9 Paper, official abstract.
- Richard A. Klein et al., "Investigating Variation in Replicability: A 'Many Labs' Replication Project", Social Psychology 45(3) (2014), 142–152. https://doi.org/10.1027/1864-9335/a000178 Paper, official abstract.
- Richard A. Klein et al., "Many Labs 2: Investigating Variation in Replicability Across Samples and Settings", Advances in Methods and Practices in Psychological Science 1(4) (2018), 443–490. https://doi.org/10.1177/2515245918810225 Paper, official abstract.










