Skip to content

‘No smoking’ signs trigger urge to light up: Communism, Marriage, Evidence-Based Medicine and the Fate of the World

Before you read the blog, please take:

General Knowledge Ethics Quiz

  1. What is the main cause of climate change?
  2. What is main cause of global poverty?
  3. Why does terrorism exist?
  4. What caused the Fukushima nuclear reactor disaster?

Write your answers on a piece of paper for reference. I will provide my answers presently and we can compare.


Brian Earp, a master’s student at Oxford University’s Department of Experimental Psychology, has found that ‘no-smoking and anti junk food adverts can be counter-productive by encouraging the behaviour they warn against’. Mr Earp asked 29 smokers to look at 25 images, some of which included ‘no smoking signs’. He found that when they viewed images of the signs they were more motivated to smoke than when they did not see the images.

Previous studies have found similar effects with alcohol. What broader conclusions should we draw from this kind of research?

You would think that a good way to stop people smoking is to tell them to stop smoking. That simple-minded approach evidently doesn’t work.

Simple-minded approaches to realising human goals have been dominant through human history. Throughout our recent history, good natured, and not-so-good natured, people have sought to influence human behaviour by inventing rituals, institutions, ideologies and a whole raft of biopsychosocial interventions. They plucked out their favourite intervention to realise some ideal. One of the most dismal failures was communism: a political system, nice in idea, but which failed utterly because it was blind to the bare facts of human psychology and motivation.

Life-long marriage, I have argued, is another ideal ill-suited to our evolutionary history and biopsychological dispositions in present day. Through most of human history, people experienced serial monogamy, relationships being terminated around 7-10 years by the death of one partner. Some societies have invented til-death-do-us part marriages for a variety of reasons, many religious. Such an institution is ill-fit for our biopsychological nature as animals.

What should we learn from this “anti-smoking messages increase smoking” research? It is a timely and humbling reminder that we are animals with natures and dispositions. We are not machines or toys that can be merely willed in one direction. To influence behaviour we need to understand the science of behaviour. We can’t just invent effective interventions by armchair or political reflection. Things we think will work, sometimes with good reason, often have the opposite effect in practice. Medicine began to recognise this about 20 years ago and this spawned the Evidence-Based Medicine movement. Many interventions which were thought on the basis of our understanding of physiology to have a beneficial effect were in fact shown by randomised controlled trials to cause serious harm and even death.

The greatest challenges, I have argued, to humanity this century arise not from external threat but from human behaviour. Climate change, terrorism, mass migration, environmental degradation, infectious disease, global poverty all result from choices we make which are a result of our psychology and dispositions. The cause of climate change is not carbon dioxide. That is the symptom of the disease. The disease is our choices, our consumerist, capitalist lifestyle. Even the Fukushima Nuclear disaster is not primarily caused by an external event – the tsunami. It is a result of the choice of Japanese authorities to locate a reactor in that place with that level of protection.

If we are to realise our goals and indeed secure the very future of humanity, we urgently need to understand scientifically why humans make the choices which they do and how we can effectively change behaviour.

This kind of psychological research is an exemplar of the most important research going on today: the science of human behaviour.

If you shove a fat man in child’s suit, the suit will burst, no matter how much you want it to fit. Indeed, no matter how “good” it would be if it did fit.

Share on

20 Comment on this post

  1. Agreed. But much of this research has already been done, so this result about smoking really shouldn't come as a surprise to anyone. Stopping smoking is a dead man's goal: something a dead person can do better than a living one. Focus on what younshouldn't do ("do not press this button") and of course you'll do more of it. Better to think *why* we want people to stop smoking, and focus on *positive* things they can do.

    Such as: always take care to ensure good air quality of those around them, take regular exercise, eat healthily, use meditation for stress relief, find ways to motivate themselves based on life goals that correspond to their values, rather than relying on feel-good chemicals.

  2. Julian,

    This is interesting stuff. I don't know if this is relevant to what you're thinking of, but simply telling people what to do (or not to do) is, of course, not the only available strategy when it comes to changing behavior. Another strategy is to design choice situations in such a way that it becomes really hard to avoid making the "right" choice, i.e., the choice that realizes the relevant good – assuming that we have good reason to single out one (set of) choice(s) as such. This, I take it, has become, if not the dominant, then at the very least a very common approach in psychology to the question of how to change human behavior, not the least within experimental economics and psychology, popularized recently by Cass Sunstein and Richard Thaler.

    Take an example that you mention: smoking. How do we help people refrain from smoking? By telling them how bad smoking is for them? If Brian is right, then that's probably not a good idea. So here's an alternative: We make cigarettes expensive and less accessible. Similarly, how do help people eat healthier? By making healthy food more accessible, and less healthy food less accessible. J. D. Trout has a good example here (in his excellent book The Empathy Gap, Viking 2009) of the health implications of accidentally bad city planning designs, which also suggests a solution: The diabetes death rate in the areas with the worst access to grocery stores is more than twice that in the areas of best access. (The correlation is stable across differences in education, income, and race.) What's the solution? Probably not simply telling people to eat better, as opposed to increasing access by (literally) changing the situation in which people make food choices.

    All of this naturally raises a whole host of issues about the viability of paternalism (including whether the relevant designs would at all be properly called paternalistic in the first place), but I take it that that's not the issue you're concerned with here.


  3. Anthony Drinkwater

    If the "discovery" that bringing an addictive behaviour to the attention of someone already subject to that addiction is likely to increase their urge is an exemplar of the most important research going on today, I think I'll stick to philosophy and double-bass playing.

    PS However, if Dr Savulescu' forms an anti-Sin campaign, I might support it from my armchair : it would be nice to eradicate poverty, terrorism and have entirely non-harmful tsunamis….

    1. Dear Mr. (Dr.?) Drinkwater,

      One more interesting aspect of the research, which is not made particularly clear in the news articles, is that the same boost in desire for cigarettes occurred whether the participants consciously noticed the signs or not, and whether they noticed their own boost in craving or not. So it's not as simple as "bringing an addictive behaviour to the attention" of a smoker. The mere presence of these signs–indeed their ubiquitous presence–may be triggering cravings outside of awareness, which, compounded over millions of instances, may indeed be something to think about with respect to public health. If you're not aware that something is going on — i.e., that a particular environmental cue is boosting your craving for some substance — then you can't do much about it. Awareness is the first step.

      Also, Professor Savulescu — I was very interested to read your take on the research, and to see how you situated it within some broader ethical issues. I do want to point out one misstatement, however. There is to date no research showing that anti-junk food messages trigger eating behaviour, and such a finding was certainly not a part of the study I presented yesterday in Glasgow. There is a study by my colleague Dr. Jennifer Harris showing that advertisements which PROMOTE junk food consumption trigger immediate consumption behaviour, but, again, no "ironic" findings in this area published in the literature. That the 'anti junk food' claim is put in quotation marks as though it's something you found in an article, or even as though it's something I said myself, is very likely to mislead your readers! I did enjoy the rest of your article.

      Brian Earp
      University of Oxford

      1. Anthony Drinkwater

        Thanks for the clarification : I didn't mean to disparage your research. I ought to know by now that one should never, ever believe a single story concerning science in the British press.
        Anthony (or to use my full academic title : Anthony)

        1. Dear Anthony — thanks for your nice reply. I like your academic title. I think the articles are roughly right about what they do report — just, in a way, incomplete …

  4. ‘If we are to realise our goals and indeed secure the very future of humanity, we urgently need to understand scientifically why humans make the choices which they do and how we can effectively change behaviour. ’Whose goals? Who are the ‘we’ that are going to change behaviour? Given the long and for the most part calamitous history of the ’science of human behaviour’, I cannot share your enthusiasm for this research. I certainly reject the notion that it and its controllers are urgently needed to save humankind from itself. Such claims have been around for centuries and in too many cases it has been the ‘research’ that has caused the harm.

    Dare I say it – calm down dear.

  5. Julian Savulescu

    Brian, sorry about misrepresenting your research. That opening paragraph was taken from the University Press Alert. Thanks for your contributions.
    Kristoffer, thanks for your helpful elaboration. Yes, the "Nudge" stuff is relevant but it is only the tip of the iceberg. It is also unlikely to have such an effect as to deal with the problems we face.
    Anthony and Keith, this is a blog, not an academic article. These blogs are meant to bring out one ethical aspect or make one ethical argument about a piece of news. If you would like to see why I think this is such an important question and this science is so important, take a look at the 6 pieces of academic writing I have published on this topic. I have been arguing we need science to improve moral behaviour. I am sorry I should have included that link in the blog. I am continuing to work on this topic as I believe it is most important in practical ethics.

    1. I just found the article you must've got the quote from — the Daily Telegraph? It does make it sound like an anti-junk food study was actually carried out! Such a finding is something we'd predict, of course, there just are no data on it yet. Thanks for your reply, and I think anything misleading traces back before your piece.

  6. Julian

    I am conversant with your work, but, as my remarks indicate, I place its claims within the context of the history and philosophy of the ‘science of human behaviour’ and find it no more convincing than earlier and other attempts to promote the ’science’ as a remedy for real or imagined treats to human well-being and existence.

  7. I agree witty Julian about the importance of this topic in practical ethics. Understanding what makes us do things, particularly when the causal relationships are somewhat counterintuitive, os crucial of we are serious about trying to create a better world.

    Keith's cautionary remarks are well-taken from my perspective, but i don't find it acceptable to rule out using the science of human behaviour to create a better future. What we *should* avoid is using that science to coerce or deceive people, and we also need to avoid using it primarily to convince people to buy things they don't really need. By contrast, using it to design better public information campaigns, and especially empowering citizens by increasing awareness of what drives our own behaviours, is to be encouraged.

  8. There's a double psychological error in the research – by assuming a simple rational actor model, it then posits a simple non-rational actor model in response.

    I'll parse this in simple terms then: the psychological behaviour you're looking for is "Raise middle finger, FUCK YOU"; or the RMFFY response.

    #1 (Most) Smokers know that smoking kills them – they also know that it is a drug, it is addictive, it is expensive, it makes their breath smell and all the negatives.

    #2 They are still rational when the exercise their free will [but they're addicts, right?! Wrong – physiological exit 2 days; rest is psychological addiction] and continue to smoke, and will indeed act in response to (mostly) patronising, insultingly low brow, uncool and sheep like motivational products telling them to 'not do it'

    #3 Thus when faced by perceived insults to their autonomy, they will, of course with (mostly) large amounts of self-referent irony, LIGHT A DAMN CIGARETTE

    This can be more easily parsed into the simple phrase:

    "Yes, I am intelligent enough to know that smoking can, and will, kill me and that its huge negative costs compared to the benefit; however, I lack the wisdom to be in a situation where I am not surrounded by grinding stress inducing feeble minded 'researchers' who don't know the first thing about the human mind, so I will continue to kill myself and at least get a momentary hit of drug pleasure from it".


    An ex-smoker.

    1. Hi John,

      You might be interested to know that half of the participants in our study did NOT consciously notice the signs at all, but showed the same boost in craving as those who did. The mere exposure had an impact subconsciously. Also none of the participants knew we were measuring craving, because the measure we used captures motivations that are expressed behaviorally but outside of participant awareness.

      So — for the people who DID consciously notice the signs, you might be right: there could be a deliberate 'fuck you' response going on in some cases. We didn't measure that. But maybe you can think of a good study design to test your idea, and I'd be eager to hear it. After all, we can all come up with theories off the top of our heads, or for the sake of a blog post comment section as you've done — but it's a lot harder to figure out if our theories actually stand up to the evidence, or apply to other people besides ourselves, or bear out what we think we know from introspection.

      In any case, even if you turned out to be right for some percentage of the cases where the signs were consciously noticed, you'd only have explained a fraction of the data. For the cases where the smokers did NOT notice the signs, your theory doesn't hold water, and doesn't make sense — and that was half of the participants in the study. I suggest there's more work to be done to understand this phenomenon.


      1. Brian

        Firstly, you'd have to prove that subliminal advertising has any basis in scientific fact. In the past 50 years, no-one has done so, to my knowledge. If you paper does so, the onus is on you to cite the previous five or so papers who sought it, and missed it. Even when the infamous "gorilla runs across the room" or (sigh) "Terrorists who view successful plane hijackings on TV are more prone to hijack planes themselves" [I won't even source that one; for someone who advised Clinton, she was somewhat.. lacking in ability. The less said about her inability to grasp that legitimising & providing 'faked' child pornography would be quickly superseded by Eastern Block 'real' content the better], we've never shown that such images have any effect…

        Or have we?

        Which leaves the old problem of certain psychological studies – those who ask the questions; it is far more likely that this is time related.

        Did you ascertain the following controls? :

        #1 Time taken to take study
        #2 Time that subject usually had between smoking (Hint: if #1 > #2 then ! Also, another small hint: smokers are unique in their habitation; i.e. each will have their own unique triggers / timing / addiction patterns, which a simplistic test such as this isn't even touching on)
        #3 Stress induced factors within the test vrs test without subliminal / visual stimuli

        If not, then the study isn't exactly scientific; given that the same subjects would have to re-take the test without the triggers to make sure it was the triggers, but… wait for it… they're prepped the second time, I'm not seeing it. But yes, you have a valid point – I'll take a proper look at it.

        General Knowledge Ethics Quiz

        What is the main cause of climate change?

        Time – CO2 released over a time period too short to allow environmental adaptation to provide balance. See Lovelock 'White / Black Daisies' or 'Calcification levels in shells of microscopic sea organisms'

        What is main cause of global poverty?

        Hierarchy [usually misogynistic]

        Why does terrorism exist?


        What caused the Fukushima nuclear reactor disaster?

        A mis-understanding of 'probability' vrs 'possibility' [the two are not synonymous].

        Oh, and super-cute! Now moderating answers. Lions, Tigers & Bears! :p

        1. Hi John,

          I can only give a somewhat cursory reply, but, in short: what we're arguing about in this comment thread is a blog post about a news item about a presentation I gave at a psychology conference. We are not, of course, talking about my actual paper, in which I do cite the relevant research on unconscious effects on behavior, assess the evidence very carefully, and address quite thoroughly many of the issues you raise (somewhat confusedly) here. My paper is on my hard drive, and I will be happy to send it to you if you give me your email address.

          With respect to "subliminal" effects, all I can say is that you've missed the last 30 years of research on the topic — unfortunately for you, because it's pretty interesting stuff and I bet you'd find it compelling. Also, even if you were right about the non-proved nature of subliminal effects, which you are most profoundly not, your criticism would be moot. My study didn't use subliminal presentation of stimuli. Instead, we showed participants images of everyday scenes in which no-smoking signs were inconspicuously embedded, but certainly visible. And we left the images on the screen for 1.5 seconds, which is far longer than the couple of milliseconds associated with subliminal effects. Some participants happened to notice the signs, and others didn't. In both cases craving went up. I'd get into more details about how this qualifies as a subconscious (note: subconscious, not subliminal) effect, but there's not room in this thread and I'm sleepy. I did find a pretty readable, publicly accessible, scientifically responsible review here: I hope you find this useful.

          Best wishes,

  9. Brian

    Thank you for the reply.

    "Carl Sagan (1987) has suggested that pseudoscience flourishes because the scientific community does a poor job of communicating its findings. To propose that we can be influenced in dramatic ways by undetectable stimuli is a remarkable claim with little scientific support, but blaming journalists for promulgating the claim absolves the scientific community from any responsibility in the educational process."

    I presume this was the point of the CSI page? If so, then would be a more focused re-direct, I presume. However, I will comment the obvious – The Mirror is never going to represent the truth of any research, as it functions to sell copy, not to promote science. "PC nanny state anti-smoking signs actually worsen what they're supposed to cure" is the immediate referent to the person writing the story; if you want my (depressing) opinion, the journalist in question has already filtered your paper through his/her <i>expectations</i> of the 'lowest common denominator' of the readership [and if the journalist missed it, his/her editor would have altered the copy downwards in any case]. MSM is all about filtered perception, even the 'heights' of the cryptic crossword.

    "With respect to "subliminal" effects, all I can say is that you’ve missed the last 30 years of research on the topic — unfortunately for you, because it’s pretty interesting stuff and I bet you’d find it compelling…Also, even if you were right about the non-proved nature of subliminal effects, which you are most profoundly not, your criticism would be moot"

    Less than 10 years ago I was taught that they didn't exist, despite a large desire for them to do so. The papers on the topic were primarily related to triggered <b>pre-stage</b> (i.e. purely physiological – nerve signals / muscle tension type data) anger / violence responses (especially to media), and the infancy of the science to manipulating a subject's responses to stimuli (children aping adults in violent behaviour). I presume that the most fertile area for papers on this will be either military (difficult to access) or Branding / Product placement at this point? (Money will drive the best research, surely?).

    If we need any pointers in the field, try this: -> skip to 3.40 and then ask me again what The Company are doing over there… (don't do so if you're worried about tinfoil being required in the future).

    In any case, the post delivered Julian's new opus (with others, of course) & the companion to bioethics (somewhat less weighty than I imagined for £100, as ever with philosophy books) so my own personal upgrade will take a couple of days. I'd be honoured to read your paper – I've set up EHC2.0 @ as an inbox.


    "calm down dear" – ouch, very Kenneth Williams there. Although, for the record, there is currently a rather aggressive anti-smoking drug on the market, which I have had direct empirical interactions with; to argue that science / cognitive altering substances aren't already being used is a total fallacy [and probably using a population of addicts as guinea pigs] . I would never attempt to sue the producers of such drugs, and I used it with full knowledge of the mechanics of what it was doing (and, being honest – made myself one of those guinea pigs as a personal moral choice)… but we're not past the days of scientific irresponsibility made famous by DDT and so forth. Short version: the 'war' for human enhancement is being waged already, which is why my interest landed here.

    1. "In sum, we have found that individual variation in an automatic motor mechanism operating at the threshold of conscious awareness is reliably correlated with GABA concentration specifically in a region of medial frontal cortex, but not in other frontal regions or parietal cortex. This result promises that we can begin to understand differences in people’s basic behavior in terms of the neurochemistry of specific brain regions."

      I'm indeed very much behind the times – and as historically anti-reductionist, probably a dinosaur. I shall troll less and apply some humility!

  10. Dear Mr Earp,

    I am a psychology student currently organising an essay entitled 'From a psychological view point, can the ban on smoking in public places be considered a total success?' and during my internet trawling I stumbled across several news reports mentioning your research around the negative effects of 'no smoking' signs. However like all newspaper reportings, they neglect to cite the research they are referring to and so I was wondering whether you would be able to send me the link to your research article; google scholar is proving pretty unhelpful! I believe that your research findings would be perfect to include in my scientific essay (referenced appropriately of course!)

    Best wishes,


    1. Hi Jasmine,

      I'm happy to share my paper with you – and a very interesting project you're working on! The articles you read were written in response to a presentation I made at a psychology conference. My actual paper is not yet in the public domain. So if you send me an email, I'll reply with the paper attached. My email can be found at under the 'about me' tab.


Comments are closed.