Drink coffee, see dead people.

January 17th, 2009 by Ben Goldacre in bad science, badscience, dodgy academic press releases, express, presenting numbers, statistics | 58 Comments »

The Guardian,
Saturday January 17 2009
Ben Goldacre

Danger from just 7 cups of coffee a day” said the Express on Wednesday. “Too much coffee can make you hallucinate and sense dead people say sleep experts. The equivalent of just seven cups of instant coffee a day is enough to trigger the weird responses.” The story appeared in almost every national newspaper.

This was weak observational data. That’s just the start of our story, but you should know exactly what the researchers did. They sent an email inviting students to fill out an online survey, and 219 agreed.

The survey is still online (in all its time-consuming glory, I just clicked answers randomly to see the next question). It asks about caffeine intake in vast detail, and then uses one scale to measure how prone you are to feeling persecuted, and uses another, the “Launay-Slade Hallucination Scale“, 16 questions designed to measure “predisposition to hallucination-like experiences”.

Some of these questions are about having hallucinations and seeing ghosts, but some really are a very long way from there. Heavy coffee drinkers could have got higher scores on this scale by responding affirmatively to statements like: “No matter how hard I try to concentrate on my work, unrelated thoughts always creep into my mind”; “Sometimes a passing thought will seem so real that it frightens me”; or “Sometimes my thoughts seem as real as actual events in my life”. That’s not seeing ghosts or hearing voices.

And of course, this was weak observational data, and there could have been alternative explanations for the observed correlation between caffeine intake and very slightly higher LSHS scores. Maybe some students who drink a lot of coffee are also sleep deprived, and marginally more prone to hallucinations because of that. Maybe they are drinking coffee to help them get over last night’s massive marijuana hangover.


Maybe the kinds of people who take drugs instrumentally to have fun and distort their perceptions also take drugs like caffeine instrumentally to stay alert. You can think of more, I’m sure. The researchers were keen to point out this shortcoming in their paper. The Express and many others didn’t seem to care.

Then if you read the academic paper you find that the associations reported are weak. For the benefit of those who understand “regression” (and it makes anybody’s head hurt), 18% of the variance in the LSHS score is explained by gender, age and stress. When you add in caffeine to those three things, 21% of the variance in the LSHS score is explained: only an extra 3%, so caffeine adds very little. The finding is statistically significant, as the researchers point out, so its unlikely to be due to chance, but that doesn’t affect the fact that it’s still weak, it explains only a tiny amount of the overall variance in scores on the “predisposed-to-hallucinations” scale.

Lastly, most newspapers reported a rather dramatic claim, that 7 cups of coffee a day is associated with a three times higher prevalence of hallucinations. This figure does not appear anywhere in the paper. It seems to be an ad hoc analysis done afterwards by the researchers, and put into the press release, so you cannot tell you how they did it, or whether they controlled appropriately for problems in the data, like something called “multiple comparisons“.

image Here is the problem. Apparently this 3 times greater risk is for the top 10% of caffeine consumers, compared with the bottom 10%. They say that heavy caffeine drinkers were three times more likely to have answered affirmatively to just one LSHS question: “In the past, I have had the experience of hearing a person’s voice and then found that noone was there”.

Now this poses massive problems. Imagine that I am stood facing a barn, holding a machine gun, blindfolded, firing off shots whilst swinging my whole body from side to side and laughing maniacally. I then walk up to the barn, find three bulletholes which happen to be very close together, and draw a target around them, claiming I am an excellent shot.

You can easily find patterns in your data once it’s collected. Why choose 10% as your cut-off? Why not the top and bottom quarters? Maybe they have accounted for this problem. You don’t know, I don’t know, they say they have, to me, in emails, but it wasn’t in the paper, and we can’t all see the details. I don’t think that’s satisfactory for a headline finding, and the first claim of a press release.

And there is another problem: putting a finding in the press release but not into the paper is a subversion of the peer review process. People will read this coverage, they will be scared, and they will change their behaviour. But the researchers’ key reported claim, with massive popular impact, was never peer reviewed, and crucially the technical details behind it are not in the public domain.

I’m sorry to see academics unblameless in this dreary situation.

Please send your bad science to bad.science@guardian.co.uk

If you like what I do, and you want me to do more, you can: buy my books Bad Science and Bad Pharma, give them to your friends, put them on your reading list, employ me to do a talk, or tweet this article to your friends. Thanks! ++++++++++++++++++++++++++++++++++++++++++

58 Responses

  1. MrSamuel said,

    January 17, 2009 at 2:51 am

    Unblameless is a brilliant word.

    In fact, this article comes as the top google result for unblameless.

    Anyway, back on topic. This is, i guess, one of those instances where the shoddy reporting of science is probably not going to harm anyone or really be considered terrible by a lot of people but is overall symptomatic of a system which is broken. Although this example won’t shake the world to its very core, the system which allows crap science into the public eye so easily is of course, a real problem.

    So yeah, well done that man.

  2. swag said,

    January 17, 2009 at 4:40 am

    As always, brilliant stuff. The kind of stuff that makes Bad Science required reading before you believe any sensationalist medical study cited in the presses and blogs.

    Thanks for being a true public service.

  3. cooldaveb said,

    January 17, 2009 at 5:14 am

    Initially I thought that such baseless alarmism might have a beneficial effect. If the population could be persuaded to reduce their coffee drinking, that should reduce cardiovascular deaths (or have I got the “coffee gives you heart attacks” meme wrong?) But then I pictured queues of teenagers outside Starbucks, hoping to get psychedelic legally. As one who has pulled his share of all-nighters for deadlines, I’m putting my money on the coffee and hallucinations both being caused by sleep deprivation.

  4. Ian said,

    January 17, 2009 at 9:32 am

    I enjoyed reading this while drinking my full caffeine coffee. The I heard David Frost on Radio 4.


  5. muscleman said,

    January 17, 2009 at 9:46 am


    The coffee and heart attacks thing is complicated. The initial studies in the ’70s linking coffee with coronary artery disease, high blood pressure and stroke got it wrong about the causative agent. Coffee was simply an association factor, the people who drank the most coffee were type A personalities who also smoked, drank, were subject to lots of stress etc, etc.

    However, in real coffee that is not filtered, like plunger or espresso there are oils that are not exactly wonderful for you. But in the totality of bad oils in your diet it is hardly significant.

    We are out of instant (fair trade naturally) so I am currently drinking a triple espresso from one of those little stove top units. I had too little coffee yesterday and ended up with a withdrawal headache. No chance today 😉

    As for my heart? I am a skinny distance runner with a damn good diet. I think I can risk the coffee.

  6. Cannonball Jones said,

    January 17, 2009 at 10:27 am

    I must say I’m truly disappointed. Up until about a year ago I was on at least 7 or 8 cups of coffee a day, often more if I worked late, and I didn’t get to see a single ghost (unless you count occasionally seeing Thatcher on telly). I demand my money back!

    Seriously though I’m glad I’ve cut down to 2 a day, almost getting to the point where I can sleep again. And to muscleman, I’m also a runner and wouldn’t dream of giving it up altogether, I need a cup before morning runs at this time of year!

  7. impostersyndrome said,

    January 17, 2009 at 11:21 am

    Muscleman, a good body of research out there to suggest that caffeine has an ergogenic effect for endurance performance, so drink up! I’m now wondering thought whether it’s just runners perceiving they’re going faster for longer and simply hallucinating a great run!

  8. phayes said,

    January 17, 2009 at 11:21 am

    “Unblameless is a brilliant word.

    In fact, this article comes as the top google result for unblameless.”

    Probably partly because it is not a word. Still, the academics surely aren’t blameless and perhaps they should count themselves lucky that Goldacre at least tries not to write like a 13 year old IRC troll when being paid for it and didn’t describe them as “t055erz” in his article. 😛

  9. Tommi Himberg said,

    January 17, 2009 at 11:31 am

    The possible benefits of drinking coffee have been in the news this week, as well.

    Researchers in Finland and Sweden found that those who drank moderate amounts of coffee (3-5 cups a day) when middle aged, had smaller risk of dementia and Alzheimers than those who didn’t drink coffee or only drank very little.

    The source is a population-based longitudinal study called Cardiovascular Risk Factors, Aging and Dementia (CAIDE).

    The full ref to the article is:
    Eskelinen, M.H., Ngandu, T., Tuomilehto, J., Soininen, H. & Kivipelto, M. (2009). Midlife Coffee and Tea Drinking and the Risk of Late-Life Dementia: A Population-based CAIDE Study. Journal of Alzheimer’s Disease 16(1), 85-91. DOI 10.3233/JAD-2009-0920.

    And a link: www.j-alz.com/issues/16/vol16-1.html

  10. muscleman said,

    January 17, 2009 at 11:34 am

    I was encouraged to find out recently that caffeine has been delisted as a restricted agent for athletics. Tomorrow morning when I go for my long run* I will get up, have a half cup of coffee, an oat biscuit and after stretching and warming up 300ml of dilute cordial and a dextrose tablet. The coffee is purely to ensure I am not troubled by a withdrawal headache.

    I’m also wondering when glucose gel packs come with added caffeine. Seems like a no brainer to me.

    *only 10.5miles alas, I am in a rebuilding phase.

  11. Dudley said,

    January 17, 2009 at 12:10 pm

    Unblameless, as a one-word example of litotes, is a good use of rhetoric to make the written word come across as speech. It violates the old principles of rhetoric but works well as an example of modern British deliberate informality.

  12. Plant said,

    January 17, 2009 at 1:03 pm

    Great piece, as usual, Ben. I very much like the way you explain capitalising on chance with your maniacal gunman metaphor. I might use that in my classes. Fully acknowledged, of course…

    With my ‘stats 101’ pedant hat on, though, I must point out that because the study is observational and appears to employ a convenience sample of self selected respondents, your statement that the key result is ‘statistically significant’ is not in fact true. There was no random selection and no known population from which the sample was drawn, so to calculate standard errors and p-values is just to construct fairy tales.

    It’s an interesting case of bad science that this practice is so widespread – and not just in psychology journals. Another example of academics not being blameless. And, OK, I’ll own up to doing it myself occasionally too.

    As a reviewer I’ve tried to make the point to some editors, but in my experience they have been unresponsive. Why, I wonder? There’s an interesting piece in there somewhere (and I use the term ‘interesting’ in its narrowest form!)

  13. emen said,

    January 17, 2009 at 1:41 pm

    Tommi Himberg,


    (Finland must have been the ideal place to do this research, because you lot drink the most coffee in the world, don’t you.)

  14. mrmuz said,

    January 17, 2009 at 3:44 pm

    Geez, never mind the ol’ Texas Sharpshooter. That question is straight out of a Scientology test. One of those ones where, if you really think about it, just about everyone has at some time or another had a moment where they heard a voice and there was no one there. You’re half asleep and the lawnmower down the road hits the right rhythm and frequencies and your noodle does the rest.

  15. T said,

    January 17, 2009 at 6:32 pm

    I love this story, I heard it on radio 1. Fair play to them somebody at the end of the report said it was a load of rubbish.

    I just bought myself a new coffee machine today actually with a timer, so my brain juice will be ready first thing.

  16. isitmedicine said,

    January 17, 2009 at 6:46 pm

    Just out of interest, why is the first sentence in the Express article as reported above “Too much coffee can make you hallucinate and sense dead people say sleep experts. The equivalent of just seven cups of instant coffee a day is enough to trigger the weird responses.”

    ie why the ‘equivalent’ of seven cups of coffee? I could understand if the first part was “too much caffeine” but surely the coffee equivalent of seven cups of coffee is… seven cups of coffee?

    Unless it’s to do with varying strengths of instant/proper coffee?

  17. The Biologista said,

    January 17, 2009 at 7:12 pm

    I love coffee, and I’m most disturbed that the rubbish science coverage in the has once again tried to paint my beloved as the bad guy. Strokes and now hallucinations. I’d be willing to bet it will eventually be linked to cancer and autism.

  18. The Biologista said,

    January 17, 2009 at 7:13 pm

    The missing noun was “mainstream”. Coffee causes haphazard typing to increase by up to 50%.

  19. Chris Wareham said,

    January 17, 2009 at 7:34 pm


    You’re right about the Finns being highly suited to caffeine studies – their coffee is immensely strong. When I was on a language course in central Finland, our coffee was brewed in a samovar like device, and at the end of the day it came out as a thick sludge. Despite this, the Finnish course tutors would still drink it, claiming it was the better than the freshly brewed stuff!

  20. peterd102 said,

    January 17, 2009 at 8:20 pm

    I heard about this on Radio 1’s Newsbeat and I was suspicious, thanks for clearing the issue Ben.

    Oh and sorry to all for my occasional acts of troll-like behaviour in the past.

  21. James-Ketteringham said,

    January 17, 2009 at 9:22 pm

    even without the misrepresentation by the academics and press – the survey itself is based on a semi-self selecting sample base of 219! Out of interest – how was the survey advertised in the email (did it mention anything supernatural – which I would imagine would draw people who believe they have seen/hurt things in like a magnet)

  22. emen said,

    January 17, 2009 at 9:49 pm

    Chris Wareham, “on a language course in central Finland”??? Haha, is there anybody on this thread who does NOT speak Finnish?

    I, too have lived in Finland, and after some time you simply need to start telling them you don’t drink coffee at all – otherwise you end up having to have one with them every hour.

    I was at the same university that Tommi is working at, by the way.

    (And now I’ve started to miss Finland…)

  23. Robert said,

    January 18, 2009 at 12:12 am

    I drink shed-loads of coffee, 10 mugs of filter coffee per day is not unusual for me. Yes, I’m a programmer. But I’ve never had hallucinations and neither do the people I work with who are worse coffee addicts than me – well so they say.

    I’m more worried about the fact that going to the toilet smells of coffee rather than any possible hallucnations.

  24. treeofpain said,

    January 18, 2009 at 12:54 am

    At the end of Ben’s Google search:

    thegrocer.co.uk has in it’s ‘Around the Papers’:

    ‘Coffee is the subject of today’s obligatory Daily Mail health scare – appa..’

    I hope that at least in the online world, if not criticism then an awareness and open recognition of the kind of rubbish some rags constantly spout spreads far and wide.

  25. Jocelyn said,

    January 18, 2009 at 2:25 am

    And I was all ready to quit the caffiene in an effort to stop my hallucinations. Thanks Dr. Ben. I’m so relieved that that won’t help me.


  26. thom said,

    January 18, 2009 at 10:40 am

    Plant: “With my ’stats 101′ pedant hat on, though, I must point out that because the study is observational and appears to employ a convenience sample of self selected respondents, your statement that the key result is ’statistically significant’ is not in fact true. There was no random selection and no known population from which the sample was drawn, so to calculate standard errors and p-values is just to construct fairy tales.”

    This seems a bit odd to me. You can calculate p values and SEs for a non-random sample. Bear in mind that perfect random sampling from a specified population is nearly always a practical impossibility. The populations in inferential statistics are abstractions. (In any case being unable to calculate a p value strikes me as no great loss.) The SE is simply the standard deviation of the sampling distribution of a statistic. It therefore tells you the variability of statistic in an infinite number of identical replications. This tells you something about how precisely you have measured the statistic. From this perspective it has been sampled from _some_ population – the tricky thing in convenience samples is to work out what population is. The nature of the sampling procedure radically influences the interpretation (specifically the generalizability of the results). This is why it is important to collect information about relevant characteristics of the sample wrt generalizability.

    In what way would non-random sampling be improved by forbidding people from calculating a SE for the sample? We’d be taking a very weak study and removing useful information from it.

  27. Tommi Himberg said,

    January 18, 2009 at 10:41 am

    emen & Chris, good to see other Jyväskylä-people here! 🙂

    There was one amusing aspect in the Finnish-Swedish coffee study. Finland might be an ideal place to study coffee drinkers, but the initial plan was to study the effects of tea, as well.

    Unfortunately, they didn’t find enough tea drinkers to get reliable results, even though they defined a “tea drinker” as someone who drinks one cup of tea per day. Perhaps that part of the study should have been done in England.

  28. muscleman said,

    January 18, 2009 at 10:54 am


    We didn’t get to Rome but our swing through northern Italy a couple of years ago brought howls of anguish from the young ‘uns when they got back and went to Costa/Starbucks with their friends.

    Fortunately we have a good little shop with an excellent range of beans and their continental roast while not quite as good as Italian coffee is not far off.

    I suspect the ‘coffee equivalent’ is in there to cover the use of things like Red Bull, iow it’s the caffeine.

  29. thepoisongarden said,

    January 18, 2009 at 11:21 am

    There is a vicious circle developing (already developed and now expanding) between academics and the media.

    Academics need to get noticed otherwise they won’t get funding for their next piece of work.

    The media isn’t interested in qualified information, it wants it straight up and down.

    Academics who rely on the true conclusions of their work soon find that all the necessary caveats mean the papers can’t get their big headlines.

    So, the academics help them along, in the press release, and you end up with the situation described in this case.

    So, who’s to blame? Is it the academics who’ve decided to feed the beast? Is it the beast itself which, increasingly, can’t afford to pay reporters for the time it would take to properly investigate these claims (i.e. actually read the paper not the press release) or is it that part of the reading public which likes to be scared by what’s in the paper and can’t wait to pass it on to others with all the qualifications removed?

  30. guthrie said,

    January 18, 2009 at 4:29 pm

    Thepoisongarden- I was under the impression that for quite a few years now it was generally the university press office that made sure press releases were sexy, not the academics themselves. This may well vary depending on location, but also suggests multiple levels of complicity; after all the university admin want the uni to look busy and important.

  31. kleptonat said,

    January 18, 2009 at 10:52 pm

    You can’t tell anything from this Psychology Department “Research” because they once again just surveyed their own students by email.

    There is no sample size and there is no response rate. There is also no control over having multiple responses from the same person or possibly from people outside of whatever the sampling population was (not that they tell you what that was exactly).

    People who have abberant behaviour (like drinking 8 cups of coffee a day- outside of Scandinavia) tend to also have high prevalences of other unusual behaviours as well. A cross-sectional (almost retrospective actually) observational self-report study with a sample size of 219 which fails to control for more than the smallest handful of confounders does not allow much in the way of conclusions.

    But to be fair to the authors they do not make strong statements about their findings. But the ensuing media storm does not exactly make them look good either.

    So basically some university students (possibly only psychology students from one university) in England who don’t smoke show an association between a psychometric measure (self-reported caffeine consumption) and another psychometric measure (maybe I hear stuff that isn’t there).

    If one were to stop thinking for a moment and try to explain this association as a sleep researcher the first thing you think of is the Narcoleptic tetrad- a set of symptoms that describe classic Narcolepsy. Whilst narcolepsy is quite rare 1/500-1000 roughly parts of the symptom cluster can be more common. The two in play here are Hypnopompic/Hypnogogic hallucinations (hallucinations around sleep onset and waking) and daytime sleepiness. Heavy coffee consumption in conjunction with hallucinations could just be Narcolepsy.

    However, since we can’t tell the real sample size in this ‘study’ we can’t know whether it’s a reasonable explanation.

  32. paulhutton said,

    January 19, 2009 at 5:37 am

    I’m a research clinical psychologist working in the area of psychosis. I don’t want to speak on behalf of the researchers involved but there are a couple of things worth noting. First of all, from my understanding of the literature I think the researchers approached the issue of coffee & hallucinations from a particular theoretical perspective which qualifies Ben’s barn shooting analogy slightly. That is, if Ben announced in advance (on the basis of previous research/theory) that he’s going to end up with three bullet-holes spaced close together then we would have more grounds for taking that result seriously. Check out previous papers by Jones looking at the role of cortisol etc.


    Secondly, I think the researchers are part of a growing movement that wish to try and ‘normalise’ the experience of voice hearing and other such phenomena. Evidence certainly favours the view that these experiences are a lot more common than used to be thought. However if they are linked to anxiety or stress then developing catastrophic appraisals such as “I’m going mad”, “I’ll lose control”, “There’s something wrong with my brain” etc might arguably lead to an increase in their frequency and intensity (given the increase in anxiety such appraisals lead to). See work by Morrison, Bentall etc. On the other hand, knowing that they might be linked to mundane stuff such as not sleeping, stress or having too much caffeine (or whatever) may reduce one’s fears. Certainly this is the approach clinical psychology in the UK take as part of cognitive therapy for voice hearing (recommended by NICE).

    For more a pilot podcast relating to this, you might like to visit:


    Anyway, I wonder whether the researchers were keen to advertise their results in order to help people – rather than scare them? Who knows & who knows what the actual effect has been, however I think a person’s intention is always important to consider in these matters.

    Finally, I’m aware that the link between cannabis use and psychotic experiences is also mediated by lots of different factors (and I bet a lot initial research was carried out on student populations!). More research clearly needs to be done .

  33. The Nameless said,

    January 19, 2009 at 1:03 pm

    Did anyone notice that you ‘fail’ one of the questions just by having a good sense of pitch?
    Whoever devised the questionnaire wasn’t tone-deaf, by any chance?;)

  34. DrB said,

    January 19, 2009 at 1:28 pm

    In fairness to the authors, press offices, this one included, often pressure you to produce some kind of ‘figure’ they can tell the press in order to make the results more ‘contrete’ and easy to communicate for the journalists.

    I avoid doing it like the plague, and when I’ve capitulated it’s usually been figures that can also be drawn from the paper. But they can be quite insistent.

  35. Synchronium said,

    January 19, 2009 at 2:34 pm

    I wonder if there’s an increased risck with my mandatory wake-up Coffee/Guarana Combo?

    My fingers are crossed!

  36. undrgrndgirl said,

    January 19, 2009 at 6:57 pm

    bad science – if you’re going to debunk stupid studies, i would suggest you do it without using propaganda…otherwise you too are just committing bad science…there is NO SUCH THING as a marijuana hangover !!!!!

  37. Ben Goldacre said,

    January 19, 2009 at 7:05 pm

    guarana is a very effective stimulant whenever i’ve tried it, similar to caffeine, legal etc, anybody come across a good review on it, pharmacology, risks?

    i love how the cannabis posse are all indignant and paranoid about the marijuana hangover. plenty of people feel bleary the next day after they get stoned. if you don’t, or want to say you don’t, or just get stoned all day and don’t notice, then i salute your somethingorother.

  38. The Nameless said,

    January 19, 2009 at 9:27 pm

    I’d guess the media response is due to an assumption that hallucination = spectacular psychotic freak-out… whereas most of the things they’re asking about are not even close– at least, I’ve had hypnogogic illusions all my life, and I seem at least *reasonably* sane…

  39. nerd said,

    January 20, 2009 at 1:52 am

    Just a small point Ben, but could you please explain the “massive marijuana hangover” thing please, as I have never heard of that one before.

  40. alibim said,

    January 20, 2009 at 3:02 am

    O-kay. I’ll use that one to counter my husband’s suggestions that I should drink more coffee as it will help ward off alzheimer’s disease. This suggestion was based on another Finnish study, funnily enough! (I wrote a bit about it myself: sci.waikato.ac.nz/bioblog/2009/01/does-drinking-coffee-lower-the.shtml)

    I didn’t know that Finland was such a hotbed of coffee-drinking 🙂 I’ll stick to my green tea…

  41. Alexa said,

    January 20, 2009 at 4:26 am

    Well I can add that this story was blasted on Australian public radio this morning. And no, no mention of it being a dodgy online survey.

  42. NorthernBoy said,

    January 20, 2009 at 10:29 am

    It is quite bizarre, and disapointing, when people scream for published data to support the most normal everyday observations, and say that it is bad science to ever mention something without backing it up in this way.

    Nerd, can you point me to the data that shows that wearing a parachute increases your life expectancy when jumping out of a plane? You see, someone claimed that it is helpful, but there s nothing in the data to support this. I cannot find any reference to a double blind trial, I can’t find any peer-reviewed studies, and I cannot even find a pilot study comparing survivability with or without a ‘chute. Is it bad science to rely on anecdotal data here? If not, why not? Why is the noting of the existence of a marijuana hangover any different?

    I really do hope that no-one actually involved in science makes this sort of petulant demand, misunderstanding that not every truth is only true once there is published and verified research available.

    I assume that you are not a user yourself (the change in my friends who are regular users is enough to convince me that the dangers are very plausible), who is just desperately trying to convince yourself that it has no negative effects…

  43. Amanda Hill said,

    January 20, 2009 at 3:44 pm

    God, you’re so naughty to recklessly click away on that survey. Anyway, babe, it’s “Imagine I am STANDING facing a barn”, not ‘stood’. I know you probably don’t care, but I have to button up your grammar when I see it lying open like that. Apart from that, yes, yes, yes, you’re absolutely right, you geek ;-P

  44. kleptonat said,

    January 20, 2009 at 10:09 pm


    Well said.

  45. emen said,

    January 21, 2009 at 10:22 pm

    “Imagine that I am stood” is correct.

  46. Amanda Hill said,

    January 22, 2009 at 12:14 pm

    His standing is continuous. He could say, “Imagine I stood”. However, Ben’s the type more likely to say, “Imagine I gave a shit” – sorry, “I am gave a shit”.

    Shall we agree to disagree? 🙂

  47. SimonCox said,

    January 22, 2009 at 1:12 pm

    I’m concerned that The Express has committed a far greater error than simply misrepresenting research. They say that too much coffee can “make you sense dead people”, which they clarify as “hearing voices”…but dead people can’t talk.

    What idiots!

  48. Lemonade Lily said,

    January 22, 2009 at 10:21 pm


    On guarana there is brief mention in ‘Adverse events of herbal food supplements for body weight reduction: systematic review’ Pittler, Schmidt & Ernst (Obesity Reviews (2005) 6, 93–111) but only few lines (and was combination preparation): adverse effects included ‘irritability, heart palpitations, anxiety and other central nervous system
    events’. Review say guarana contains caffeine and used as weight loss aid….which you don’t appear to need…….

    Another paper is by Subbiah ‘Guarana consumption. A review of health benefits and risks’ (Alt Complement Ther 4,212–213 (2005)) – I’ve not read

    On coffee – at weekend I came across a website encouaging curious bathroom antics with coffee via a different orifice (www.natureworx.com/downloads/The_coffee_enema-1.pdf)’Circling the feet and hands helps to get you through waves of peristalsis’


  49. emen said,

    January 22, 2009 at 10:33 pm

    Amanda, of course his standing is continuous! 🙂

    “Imagine that I’m stood” is an informal way of saying: “Imagine that I’m standing”.

    NB “stood” is the past participle form here, not the past form
    so your example sentence “I am gave a shit” should be “I am given a shit”

    however, “give” is a transitive verb, whereas “stand” is intransitive

    therefore “I am given a shit” isn’t correct as an informal way of saying “I am giving a shit”, because “I am given a shit” is a passive construction, meaning “somebody gives me a shit”

    aren’t we boring

  50. thom said,

    January 23, 2009 at 1:09 am

    Plant ‘The sampling variance that is estimated by the SE allows you to say how precise your estimate of the population parameter is, not the statistic, as you suggest.’

    Fair point – I meant parameter.

    ‘Given we don’t know the population, what use is any kind of estimate of ‘precision’? Precision of the estimate of what?’

    As you say this is arguable, but the population is defined by the sampling procedure (not by what the researchers say they sampled). I think its OK to think the sample as a sample of some population – and in this case the sampling procedures makes it very hard to say anything useful about what population. This is a fairly standard interpretation (e.g., in Hays, 1973, Statistics for the Social Sciences). More importantly, if your argument is taken at face value it means we can virtually never calculate SEs, CIs or p values – because we can never ensure every member of the population has an equal probability of selection during sampling. Scientists also rarely work with defined finite populations (and assuming an infinite population is slightly conservative in that it increases the width of a CI).

    Reporting a mean without a CI (or SE) is nearly always worse than reporting it with the CI. Just giving the point estimate can give an impression of false certainty. (A common problem with reporting just point estimates and p values). Anyway, the problem here was more with the press release than the paper (a short report in a specialist journal).