Tuesday, April 5, 2011

Comic 882: Signifying Nothing

[I almost forgot to mention that this is a guest post by "Ann Apolis," noted transvestite. -Ed.]

Greetings, cuddlefish and sucksters alike. I am Ann Apolis M.D., here to cast a surgeon's eye (someday I hope to find out which surgeon) over the latest xkcd. Is strip 882 a simple case for outpatients, or is its lack of humour inoperable? Do I inform the next-of-kin with a smile on my face and a couple of jokes or a sombre expression and a comforting voice? Will the tenuous medical analogies continue? I'm as eager to find out as you.

Here it is:

significant
[Comic title: Significant; Alt text: 'So, uh, we did the green study again and got no link. It was probably a--' 'RESEARCH CONFLICTED ON GREEN JELLY BEAN/ACNE LINK; MORE STUDY RECOMMENDED!']

Once again Randall takes it upon himself to point out how stupid people are [though not his esteemed readership of course, who are all intelligent enough to laugh at the idiots in this comic]. Here it's sloppy statistics. Once you notice the 0.05 and the fact there are twenty tests, the final joke [such as it is] because rather obvious and Randall doesn't deviate from the obvious course in presenting it; we clump, clump, clump to the finish line with weary inevitability.

It's not completely awful; while the Minecraft references are a bit 'GOOMH RANDALL I PLAY MINECRAFT TOO', they are at least attempts to include little jokes on the way to the big joke, and this is something Randall doesn't usually do. [In a similar vein, I was hoping there would be some amusing colours in the TWENTY PANELS OF DOOM, but no dice.]

But the overall impression this comic gives is of an amusing illustration in a textbook, the kind that would get you to go 'Hey, I like this textbook, it's kinda funny'. Except this isn't a textbook, this is The Most Popular Webcomic On The Internet (TM), and this kinda thing doesn't cut it. It's badly drawn (come on, Randall, I don't blame you for copy-pasting the twenty panels, but you could at least have made the heads join up in the original), it's patronising, but most of all it is dull. Dull dull dull dull dull.

And maybe I wouldn't mind, if it would end here. But it won't. For the next week this is going to be passed around with a self-congratulatory pat on the back, a smile of "Aren't newspapers stupid? At least we know how science works". And for years to come any time a medical study is mentioned someone is going to post "ah, but did they take THIS into account ". And everyone will go "Hahaha, that's right, that is a thing that happens with statistics". And the myth that xkcd is funny and special and the cream of the internet will continue.

ORWELLIAN WATCH: Originally the comic said "green jelly beans" in panel 2 (http://www.webcitation.org/5xjact3OU) which kinda makes the joke stupid. Quality control at Randall Towers is as high as ever.

159 comments:

  1. Naturally since I ended the post by making a dig at Randall's lack of quality control, this post originally said it was talking about Strip 881.

    how... ORWELLIAN

    ReplyDelete
  2. I love how little it seems that the general population understands statistics...

    ReplyDelete
  3. Also, the idea that 20 test at 95% confidence each has a 100% chance to yield one false result is not at all correct. In reality, there's only a 64% chance (1-(.95^20)) that there would be one (or more!) incorrect results. Likely, but a third of the time you'll be just fine.

    ReplyDelete
  4. Yes, of course newspapers check the actual results of a journal paper/experiment and then report it. It's not like they rely completely on press releases made by the scientists themselves (or a specialised company if they're rich)! And of course scientists don't oversell marginal results just to get published! Only the media does that!

    Also, fuck image spamming. Just fuck it.

    Captcha: Fliery. The dragonfly had fliery breath.

    ReplyDelete
  5. @ Anonymous, the comic doesn't claim that there is a 100% chance of yielding one false result.

    ReplyDelete
  6. Hey, I'm not even a scientist and I found a link between jelly beans and acne: jelly beans are high in sugar.

    You can try and be pedantic about that "myth," but factually once you've had more sugar than your system can handle it's going to trigger the relevant hormonal reaction. I'm going to talk about this on the xkcd forum and madness will ensue.

    ReplyDelete
  7. You know, that's what the Bonferroni correction is for. Multiple tests.

    ReplyDelete
  8. @12:20 Yeah, that actually really is a common misconception.

    ReplyDelete
  9. @12:49 Passive aggressive attacks? COMMON MISCONCEPTIONS?
    Go back to the xkcd forums, you cuddlefish.

    ReplyDelete
  10. Rob, get a life yah? stop wastin time posting the same old shit about the same old shit

    ReplyDelete
  11. "It's not like they rely completely on press releases made by the scientists themselves"

    Kitten you are a fucking moron- a majority of the time the press releases the newspapers tend to just flat out ignore or misinterpret.

    http://www.badscience.net/2011/03/why-dont-journalists-link-to-primary-sources/#more-1997

    there you go. References and everything.

    Also I'm not sure I understand this comic- mostly because I also don't understand statistics he's using here.

    surely if you're testing for correleation between jelly bean colours and acne you'd preform a seperate closed study for each colour?

    tl;dr, I'm also a fucking moron- take my hand and guide me through the steps of Randall 'ihaveadegreedontchaknow' Monroe

    ReplyDelete
  12. Charles Augustus FortescueApril 6, 2011 at 3:06 AM

    What is the objection to the heads not joining up to the bodies? I for one do not find it to be a problem.

    ReplyDelete
  13. You don't want the correlation to be happening by mere chance, so you built a confidence interval around your result, usually by assuming that the errors are normally distributed. If you find a positive correlation and 0 is not in your confidence interval, you can be "almost" sure that the correlation is true. The question is: what does almost mean. You can build a confidence interval at the 20%, 10%, 5% level. The most commonly used is the 5% level. A confidence interval at the 5% level means: there is a 95% chance that the true correlation lies in that interval, and a 5% chance that it does not.

    A 5% chance is 1 chance over 20. So if you do 20 experiments, it can happen that you find one positive correlation, even if there is in fact no correlation. That's the joke. And then, the stupid journalists go like "there is a correlation", WHEN IN FACT THERE IS NOT!!! See, that's why it's funny.

    I'm a bit disappointed, he could have done a "correlation is not causation joke." Teenagers have acnes, teenagers eat jellybeans, so there is probably a correlation, but maybe not a causation.

    In fact, the problem with jellybean studies is that the results are distorted by one huge outlier, Rob, who eats half the US production of jellybeans and has an impressive density of spots and pimples on his face, shoulders and belly.

    ReplyDelete
  14. orite I thought it was some kind of mathematical quirk related to T-tests or something that only applied for 1/confidence interval.

    when aparrently it's just that as you preform a number of studies the increasing number of studies increases the probability of a false positive that lies out with your confidence interval.

    like uh as anon said %64 chance of a false positive for twenty (essentially) identical studies- I get it now.

    ReplyDelete
  15. does Randall get every bit of math he tries to make a comic about wrong?

    ReplyDelete
  16. Why did they do the yellow test twice?

    ReplyDelete
  17. Why did the scientists test yellow ones twice?

    ReplyDelete
  18. No, the second yellow is now mauve.

    Captcha: revinsi, it's clearly going for revision.

    ReplyDelete
  19. ^
    'Cause all the things he does, they are all yellow.

    ReplyDelete
  20. One thing I also don't understand is, if the joke here is "haha newspapers are stupid", why doesn't the headline read "GREEN JELLY BEANS CAUSE ACNE!" instead? What kind of newspaper would post the heavily unlikely and obscure "GREEN JELLY BEANS LINKED TO ACNE! 95% CONFIDENCE" headline, Randall?

    ReplyDelete
  21. Oh, look at that. Randall "I get no do overs" Munroe did TWO corrections. Maybe he should, instead, revise his comics twice!



    Also, my comment got eaten by the xkcdsucks comment box again(I commented fine in reboot, so I have no idea what the heck is going on with this box), but I really missed the point of Randall's punchline, too. The newspaper just said what the scientists said: "Green jelly beans linked to acne". Sure, that isn't headline material, but that's hardly the point.

    ReplyDelete
  22. I really missed the point of Randall's punchline, too. The newspaper just said what the scientists said: "Green jelly beans linked to acne". Sure, that isn't headline material, but that's hardly the point.

    That was my thinking, too. In a scientific journal, if your statistical analysis passes the t-test, damn right you're going to say there's a correlation.

    ReplyDelete
  23. "Kitten you are a fucking moron- a majority of the time the press releases the newspapers tend to just flat out ignore or misinterpret."

    No, the majority of the time the papers will just copy the press release. Sometimes they will fuck it up for a story.

    The majority of the time, the scientists behind the study will oversell their results in order to secure funding, especially those scientists putting out a press release. Blaming just the media for outputting dodgy science is the sort of fuckwittery I'd expect from an xkcd fanboy.

    http://www.badscience.net/2009/05/dodgy-academic-pr/

    I can quote Ben Goldacre too.

    ReplyDelete
  24. What a thrilling comeback after such a lackluster performance over at reboot. You can only use CRUISE CONTROL to keep you out of the heat for so long though.

    But the xkcd fanboy harassment here has a superb quality to it.

    Still though ...

    Reboot does a Redux countdown: 14

    ReplyDelete
  25. I wanted to add that it's way harder to prove an absence of correlation than to prove a correlation. You can basically prove all the absence of correlations you want by having a very low sample size.

    The problem with this comic is that it looks like another "science journalists don't get science" joke, when it is in fact the story of 2 stupid scientists.

    ReplyDelete
  26. Kitten, linking to supporting evidence in an online argument is cheating.

    ReplyDelete
  27. except that relying on press releases is the hight of ignorant fuckwittery no matter which way you look at it?

    ReplyDelete
  28. Yes, press releases are completely useless. People should just man the fuck up and read the journals directly.

    ReplyDelete
  29. Also, as usual, Zach Weiner did this before and much better. A year and a half ago.

    And Munroe already touched on it in 465 a year before that.

    And Carl touched on Munroe too. Actually some guest poster named Esteban touched both of them and it was terrible.

    I'ts not the fact that it was just done before, but that it was done better in 465 (although there is much room for improvement, and 465 has its own problems).

    The downward spiral continues. There are highpoints for sure, but they are much more encumbered with shit like this that I will have to read over and over again whenever someone mentions how shitty science journalism is.

    ReplyDelete
  30. I was going to comment on the post for 465, but apparently Rob has some fascist "approval" thing going on there so I'll try here:

    Wow, it's kind of shocking that there was a time where Rob could write a clever, serious (if still totally wrong) post instead of the frabro misogynistic "hurf durf he needs to get laid" stuff he posts now.

    ReplyDelete
  31. Hurpa Derp, imma cuddlefish that doesn't understand that 465 was a guest post at a time when everything was posted by Carl.

    Even though IT WAS FUCKING SPELLED IN THE LINK I JUST FOLLOWED.

    Two levels of wrong everybody!!

    As for why there is moderation on older posts I gots one word for ya: assturds.

    ReplyDelete
  32. Yeah, you're such a frabro, Rob. Quit being a frabro already

    ReplyDelete
  33. That's the real problem here. Not that Rob's dysfunctionally large, an ass, or a troll -- it's that he's a frabro.

    ReplyDelete
  34. ...okay, I give up. What the fuck is a frabro?

    ReplyDelete
  35. Oh yes, how could I have possibly missed that it was back when Carl was the main poster. And of course it was a guest post, which means it was written by...not Carl? So of course I'm an idiot for thinking it was written by Rob. It's all so obvious now.

    ReplyDelete
  36. Frabro? Google tells me it's rule 34 on Brontasauruses [or maybe they're fighting, I don't know]

    ReplyDelete
  37. what is it with people and calling me a misogynist on here? is making fun of Randy for his creepy obsession with a woman misogynistic now?

    ReplyDelete
  38. I'm pretty sure that comes part and parcel with being a frabro.

    ReplyDelete
  39. Okay, first of all, Fake Mole, stop trying. Unless you can replicate that OpenId icon next to my nick, you won't fool anyone.

    Also, that was one lame fake comment, hm?

    I forgot whatever else I had to say. So whatever.

    ReplyDelete
  40. @Rob
    Not a misogynist. A FRABRO misogynist (oooooh)

    ReplyDelete
  41. What's great, is that if the scientists are actually practicing their experiments in a proper understanding of statistics (hint: they do) and they're claiming 95% confidence intervals on each of those tests, the conclusion in the final panel is totally accurate! The only way that the error would be part of an expected exception would be if each of the jelly bean color tests were the result of one larger experiment, which if they're testing the conclusion they claim to be testing ("color X jellybean is positively correlated with acne") they'd never do! which means Randall is full of shit!

    I mean, the paper gets it wrong with the "only 5% chance of being wrong" thing, but the notion that such a test, if actually describing what the scientists claim it does, wouldn't be statistically significant, terribly misunderstands basic experimental statistics!

    ReplyDelete
  42. a 95% confidence interval does not mean that it is 95% probable that the true parameter is inside the interval. The population parameter is fixed, so that probability is either 100% or 0%.

    a 95% confidence interval means that, if the procedure that generated that interval was repeated many times, we'd expect 95% of those intervals to contain the true population parameter.

    Incidentally, this undercuts Randal's entire argument, since those 20 tests were all independent of one another.

    ReplyDelete
  43. Actually, that's exactly why it isn't.

    ReplyDelete
  44. undercut, i mean.

    ReplyDelete
  45. Yeah, Randall fails at statistics. Like others have said, if these tests were performed independently- and I don't see why they wouldn't be, if the scientists are testing for a correlation between color and acne- the result is definitely statistically significant. And even if it were all part of one larger test, the chance of an error isn't 100%. I know the comic doesn't come right out and say the chance of an error is 100%, but it REALLY seems like that's what he's thinking, considering he had 20 tests at 95% confidence. It seems like he's thinking 5% chance of error x 20 tests= 100% chance of error, which is totally wrong.

    ReplyDelete
  46. Should point out: when I say "Once you notice the 0.05 and the fact there are twenty tests, the final joke because rather obvious", I don't mean I genuinely think that something with a probability of 1/20 happens exactly once every twenty times you try it; I mean that for the purposes of a comic on the internet showing it happening once in twenty times is the most obvious way to get your point across.

    ReplyDelete
  47. Animated gif here that flashes between the mistake and corrected versions of this comic.

    http://forums.xkcd.com/download/file.php?id=29068&mode=view

    Notice how he changed 'coincidence' as well, rewriting the word and adding an exclamation mark.

    ReplyDelete
  48. It's too late, Ann, it's too late. Your horrifying error indelibly stains your name, now and for evermore.

    THAT IS THE PRICE YOU PAY FOR GUEST-POSTING.

    @Jon - Does Randy ever post in his own forum? He obviously reads them, because he uses them as if they're his editors.

    ReplyDelete
  49. Anon 6:03 PM

    So I guess Rob occasionally likes to pretend to be a somewhat dumber man named Esteban?
    It all makes sense now.
    Tacos.

    ReplyDelete
  50. Chad Fourtwentington Raperson III, IAF PresidentApril 7, 2011 at 4:20 AM

    Hi, I'd like to clarify something here. The dude known as Rob has had his membership suspended on account of not paying his dues ever an being an enormous fatass. Nothing he says or does should be taken as representatives of frabros as a whole. Even we have standards that don't involve skin color, after all.

    Thanks,
    Chad Fourtwentington "Stoptheobamanation" Raperson III, Esq.
    President
    International Association of Frabros

    ReplyDelete
  51. Mr. Owl, how many times does it take for Randall Munroe to fuck up basic science for his fans to stop gagging themselves on his cock?

    ReplyDelete
  52. Come on Chaos and Anon@11:37, Randall may make plenty of statistical mistakes but he's not that dumb. I can't believe you're making me stick up for him.

    Ann gets it; he's showing something happening 5% of the time, not claiming that something with a 5% chance must happen once every 20 times.

    Well, I'm off to the casino. I've got a great roulette strategy. I'll just watch for 37 rounds and remember the results, then place a bet in the 38th round. Whatever number doesn't get hit in the first 37 rounds is guaranteed to be the winner in the 38th round.

    ReplyDelete
  53. anon 7:25 this is how randall monroe spends his T shirt money. when he doesn't get his result he throws a geekfit and calls the game rigged- but he always returns for another play

    ReplyDelete
  54. Guys, I have a confession to make. 27 years ago I shat out some bubbly primordial soup and out crawled a putrid abomination that even a puppy couldn't love. I named it Randall, which means "festering pile of dickshit" in Latin. He had a tendency to crap out shreds of fuckpaper and call them comics, a practice I tried to discourage by reminding him that he came out of my ass and was therefore incapable of accomplishing anything that wasn't shameful at best. Failing that, I gave up and sent him off to whatever school would take him.

    Of course he sucked at that too much to go anywhere with it, so he naturally sunk back into his old habits of labeling his shit and showing it to people. Somehow he got his own website to share his repulsive habit with strangers.

    Not a day goes by I don't feel remorse for not putting it out of our misery when I had the chance.

    Sorry, guys.

    ReplyDelete
  55. Hey guys? There's some horrible scandal about Randall going on, but I can't find the details. Somebody fill me in.

    ReplyDelete
  56. A Randall scandal eh?

    Sounds improbably juicy. We can probably make up details and present them as fact. That would be a completely responsibly ting to do.

    I heard it involves the real Megan, an iPad, a plunger and some Python. He might get sent to the clink over this one.

    ReplyDelete
  57. Vaginal secretion

    I heard the scandal was that he spent years working on a shitty webcomic, forgot how to science, and is no longer capable of making it through grad school or getting a physics related job. I also heard that his fanbase is steadily drying up as the comic's quality deteriorates, but slowly enough to keep him trying to save it. More specifically, I heard that the comic will die on Randall's thirtieth birthday, and although he'll realize he's spent his twenties without gaining real work experience or building savings, he'll publicly claim that he had a blast and regrets nothing. At night though, you'll often hear the sound of sobbing through his door.

    Just a rumor.

    ReplyDelete
  58. responsibly ting

    captcha: expend
    randall is pretty expendable

    ReplyDelete
  59. I raped Randall in the ass after I beat him down in a fight. The cum's probably still in there somewhere.

    ReplyDelete
  60. Penile burninationApril 7, 2011 at 12:19 PM

    well I raped randy in his vag

    ReplyDelete
  61. Don't scientists doing a series of studies usually set a lower p-value for just this reason?

    Of course, this does happen with unrelated studies, with all the studies that go on, but there's no way around that.

    ReplyDelete
  62. fact: randall's mother reads this blog. used to, anyway

    seriously

    ReplyDelete
  63. Reverse ejaculation

    Sorry Randy's mom. All posts here are made by one person: me. I was friends with Randall in high school and I run this blog as a joke. He knows about it, it's cool.

    ReplyDelete
  64. You probably had her convinced until you said he had a friend in high school. Randy's mom knows better.

    ReplyDelete
  65. I heard Randy's mom is actually Carl, and Randy just found out.

    ReplyDelete
  66. Did..... did Randy just copy and paste for those 20 panels?

    Oh god.

    ReplyDelete
  67. rectum unclean

    Why 1:14? If he drew 20 slightly different shitty drawings I'd still be unimpressed with the art. He may as well embrace his disartistry.

    ReplyDelete
  68. Well, at least it would be less FREAKING BORING if the 20 shitty drawings had more differences than a single word.

    ReplyDelete
  69. OR, he could just have one panel where they list all twenty colors.

    ReplyDelete
  70. Yeh, cos that would of looked SO awesome.

    The copied panels are probably the only bit of comedy in this comic. Repitition = humour.

    ReplyDelete
  71. Dear anon 8:13,

    With respect to your comment,

    "a 95% confidence interval does not mean that it is 95% probable that the true parameter is inside the interval. The population parameter is fixed, so that probability is either 100% or 0%."

    Let's go Bayesian for a while: a 95% confidence interval does indeed mean that, given the results of the experiment, the true parameter, which is indeed fixed, but whose true value I do not know, lies with a 95% probability in the confidence interval.

    It is a subjective probability, depending on the current state of my knowledge. But most of the probabilities we deal with are subjective.

    Let's assume I bake 3 birthday cakes. One of them contains a 3DS, the other one a naked lady and the third one a naked Rob. Then I select a cake at random, and offer it to you. What is the probability that it contains naked Rob? (the cakes are all huge, so that you cannot go by the size; moreover, Rob has been muzzled, so that he cannot eat the birthday cake from within).

    You might make a smartass move and say that this is either 0% or 100%, depending on whether naked Rob is in the cake or not. But you don't know what's in the cake. Hence your subjective probability is 1/3.

    ReplyDelete
  72. I enjoyed this review. Keep up the good work.

    ReplyDelete
  73. Randall fucked up the signs on everything in this comic.

    We reject the null hypothesis (meaning that the data isn't just a fluke) if the p-value is LESS than .05, not if it is GREATER than .05. A p-value is the chance of it being a fluke.

    Randall says "No link between jelly beans and acne, p>.05" which means that the probability of it being a fluke is MORE than 5%: we usually just say "It's maybe a fluke, let's test it a little more and try to reduce that p-value, or maybe scrap it."

    The fact that he hasn't corrected it and it is still Thursday means that he and everyone else who faps to xkcd is retarded.

    Also, I'm ashamed of everyone here for not seeing that.

    ReplyDelete
  74. I am also annoyed at everyone for failing to post hypocritical, self-righteous judgments about things you don't understand.

    Shame on you all.

    ReplyDelete
  75. It's probably worth explaining to Rochambeau that p-values and confidence intervals aren't the same thing.

    ReplyDelete
  76. @6:41
    Slobbery cock

    nothing is ever worth explaining because nothing ever matters

    ReplyDelete
  77. Why are there two doctors in the newest xkcd? The brunette one (OF COURSE THEY ARE BOTH WOMEN) serves no purpose other than to set up the punchline and just stands awkwardly for the rest of the comic.

    Also this is probably Randy's worst case yet of "people levitating above furniture." And it's really, really hard to do a joke that relies entirely on reaction shots and over-the-top emotions when NONE OF THE CHARACTERS HAVE FACES.

    ReplyDelete
  78. Is it really that hard to connect a head to a body?

    ReplyDelete
  79. No, 2:33. REPETITION = UTTER BOREDOM.

    CAPTCHA: redos. Randall doesn't get them, he says.

    ReplyDelete
  80. There are not two doctor, the one wearing the white doctor coat is a doctor and the other one is presumably the guy's significant other (Megan).

    I think the joke in 883 is ok, but the execution is horrible as usual. Maybe he could have drawn what the guy was imagining (and then he just eliminated the last two lines of awkward dialogue).

    For a funnier joke about the pain scale see http://hyperboleandahalf.blogspot.com/2010/02/boyfriend-doesnt-have-ebola-probably.html

    ReplyDelete
  81. Whenever a doctor has asked me to rate pain from 1 to 10 they said that 10 was the worst pain I've ever EXPERIENCED. Not IMAGINED. Probably for this exact reason.

    ReplyDelete
  82. Also on the topic of better pain scale related jokes, this from Scrubs http://www.youtube.com/watch?v=UOL7JdzGa64

    ReplyDelete
  83. I can imagine my every nerve being afire with agony without even battering an eyelid at the thought. His imagination is normal enough, but he must be a very emotionally fragile person.

    ReplyDelete
  84. 9:37, pain experienced is equally dumb, they have no knowledge of the person's past experiences. i myself would rate a stubbed toe as about an 8 based on that scale.

    what i've seen doctors do is hit patients harder and harder until they say, ok, your pounding hurts more. that way you get an accurate measurement.

    ReplyDelete
  85. Fuck me, 883 could have been a moderately funny comic. FUCK

    Lose panel 3, we are not plankton, we remember the question from ONE PANEL AGO.

    Lose the post-punchline dialogue. It not only comes after the punchline, it's painfully unfunny. In fact, lose everyone in the last panel, focus on the guy.

    This would have been done well by Nedroid.

    ReplyDelete
  86. 10:15, How would you rate the pain, from 1 to 10, where 10 is the worst pain you can imagine?

    ReplyDelete
  87. I think the joke works better with the man repeating the question. It's mainly those two chicks chattering away at the end which blows the whole deal. It's just another "woo, so quirky!" indulgence, and it's sickening. It does cue the forum goers to scramble over themselves to prove that they're that quirky too, however, so I guess it serves xkcd's purpose well.

    ReplyDelete
  88. Gee, randall is making a lot of medical comics lately. I wonder what prompted this one?

    ReplyDelete
  89. "How would you rate the pain, from one to ten, where ten is reading Abstruse Goose?"

    "Reading Abstruse Goose?"

    ...

    "One."

    ReplyDelete
  90. "
    With respect to your comment,

    "a 95% confidence interval does not mean that it is 95% probable that the true parameter is inside the interval. The population parameter is fixed, so that probability is either 100% or 0%."

    Let's go Bayesian for a while: a 95% confidence interval does indeed mean that, given the results of the experiment, the true parameter, which is indeed fixed, but whose true value I do not know, lies with a 95% probability in the confidence interval."

    Noooo it doesn't. If I had taken a Bayesian approach then I could have constructed a 95% interval which would mean that. But I didn't. Now of course technically I could construct an interval which will be identical to that confidence interval by taking a noninformative prior, but statisticians usually do not.

    ReplyDelete
  91. Rob, I sent you a fucking guest post. Fucking read it or some shit.

    ReplyDelete
  92. I thought the dark-haired woman was his mom.

    Anyway, after he picked "one," I anticipated maybe a moderately funny idea of what was so horrible (at least Justin Bieber), but no, Randall's too lazy to actually think of something.

    ReplyDelete
  93. HAHA THAT'S SO FUNNY. HIS IMAGINATION IS SO FUCKING AMAZING HE MUST BE IMAGINING SOME KIND OF HORRIFYING PAIN. I DON'T KNOW WHAT IT IS, BUT IT MUST BE HORRIFYING. AND PAINFUL. OMG LOL

    ReplyDelete
  94. And here we see Randall demonstrating yet again the "xkcd school of humor". Principle #3: let the reader make it funny for you!

    ReplyDelete
  95. "I thought the dark-haired woman was his mom."

    Wait, you're saying Megan ISN'T Randall's mom?

    ReplyDelete
  96. Reboot does a Redux countdown: 13

    ReplyDelete
  97. I thought Megan was his grandmother.

    ReplyDelete
  98. So today I was sitting in a meeting with colleagues and everyone was passing around strip 26 laughing at what a good joke it was and I just had to bite my tongue instead of reminding them that there is no joke in that strip and then someone went on about how the transform sketched out is probably of a real "meow" since Randall is SO smart and ALWAYS checks all of his facts and oh god I could only just sit there with a blank stare on my face because I WAS THE ONLY ONE WHO THINKS THAT RANDALL IS A HACK WHO WRITES DOWN INSIPID REFERENCES TO NERD STUFF AND PASSES IT OFF AS A JOKE TO OTHER UNSOPHISTICATED RUBES HOLY SHIT MY LIFE IS NOT WORTH LIVING

    ReplyDelete
  99. Just let it all out. You'll feel much better.

    It's like holding in diarrhea--you just can't/shouldn't do it.

    ReplyDelete
  100. http://goatkcd.com/883/sfw

    AHAHHAHAHAHAHAHAHAHHAH
    OH GOD
    OH MY GOD THIS HAS TO BE THE BEST SO FAR

    ReplyDelete
  101. is this blog going to shut down when Randall announces that he's terminal

    ReplyDelete
  102. It's proven with p > .05 that randal writes for goatkcd

    Captcha: anker. A kanker sore ranks about a 5 compared to an anker sore

    ReplyDelete
  103. @12:51

    Good god I hope not. I'll try to keep it going as xkcdsucks:revamped (for at least 30 more comics anyways).

    Captcha: rugma, what is left over when you wipe off your uncirc'd member on the carpet.

    ReplyDelete
  104. Don't lie, Capn. You can't even be bothered to write out the word 'uncircumcised', let alone an entire blog post.

    ReplyDelete
  105. All I do is lie and produce fragmente

    ReplyDelete
  106. Well, this is interesting.

    http://www.cnn.com/2011/US/04/08/particle.physics.tevatron/

    Headline: "Is it a new particle, or just a fluke?"

    Note that the null hypothesis is presented in the headline. I can't be bothered to track down the actual research paper, but it appears CNN is actually talking about the p-values. Apparently the researchers have a 0.001 p-value, but are holding out for a 0.000001 p-value before they really get excited.

    I'm not sure if this supports or invalidates Randall's claims about the media, but I'm leaning towards invalidates. CNN is actually taking this story with an appropriately sized grain of salt.

    ReplyDelete
  107. Actually, all of the reporting on this (Google News for "Tevatron") seems to be pretty excellent vis a vis statistics:

    MSNBC:
    "New subatomic particle: real or anomaly?"

    BBC:
    "Accelerator hints at new particle"

    Of coure, not all of the headlines are appropriately skeptical. Any surprise that Fox News does poorly?

    Fox:
    "Mystery Atom Discovery Has Physicists Abuzz"

    Captcha: aphoti-yet another spelling for "fish"

    ReplyDelete
  108. there is a disproportionately high number of british statistics students in this thread how can you explain that i bet you can't 'cause statistics doesn't work huh?????????

    ReplyDelete
  109. Recommended reading for all:
    http://norvig.com/experiment-design.html

    ReplyDelete
  110. to whomever says they sent me a guest post:
    I didn't get it and it's not in my spam folder. try again.

    please send it with the subject line 'I HAVE MADE A TERRIBLE MISTAKE ABOUT COMIC 883', to the email address in my blogger profile.

    ReplyDelete
  111. Oh fuck. I sent it to the wrong place.

    Fuckfuckfuck. K I sent it.

    ReplyDelete
  112. Yes... I was very surprised to receive such a disturbing piece of electronic mail, and assumed it was a mistake. What you do during your free time is none of my business as your research supervisor, but I am concerned about your loyalty to science, mental health and ability to pursue a PhD in our university.

    Xkcd is a funny webcomic, I must say, and your violent review took me aback.

    You must send me your referee report by Monday, and do not forget about the Wednesday seminar (again).

    ReplyDelete
  113. "And the myth that xkcd is funny and special and the cream of the internet will continue."

    By definition, it is funny if one gets a chuckle out of it. You need to get your brain wrapped around the nature of humor, which in part is that it's rather subjective, specialized. If you don't think it's funny, then don't laugh. You're free not to.

    ReplyDelete
  114. "And here we see Randall demonstrating yet again the "xkcd school of humor". Principle #3: let the reader make it funny for you!"

    See my earlier comment. The reader ALWAYS has to play a role in making it funny. If you've ever laughed at someone else's misfortune, then surely you don't think that poor soul thought it was funny. It was you who laughed.

    ReplyDelete
  115. one of the dumbest mistakes a human can ever make is assuming that his definition of a word is the "correct" one. much like someone who thinks that a funny thing is defined by people laughing at it

    ReplyDelete
  116. "much like someone who thinks that a funny thing is defined by people laughing at it"

    The audience always play their individual roles in these things. Literature, movies, etc, aren't one-way streets. There's no such thing as a joke which is funny when no one is there to laugh at it. Humor is emergent, contingent and provisional. It's not elemental or irreducible. Presuming to police people's ideas of what they think is funny is like presuming to know which faith is the one true religion (and all others are infidels). You gotta be awfully anal to fret over a comic strip that you're not even required by law to read.

    ReplyDelete
  117. that's your definition of humor, yes. the problem is that your definition is not the "correct" one. thanks for playing though!

    ReplyDelete
  118. "that's your definition of humor, yes. the problem is that your definition is not the "correct" one. thanks for playing though!"

    What you're saying, then, is that >you< have no choice but to laugh, because Funniness is infused into a joke by the Stars in the Heavens; and as a corollary, that people who laugh at stuff which you don't are mentally ill. Your profile says that one of your few interests is "pretentiousness". Oh, the irony.

    ReplyDelete
  119. actually I haven't defined humor at all. I'm just saying your definition is wrong. you are operating on the assumption that everyone shares your definitions; this is patently false, and makes you a fucking moron

    ReplyDelete
  120. I'm not defining humor either per se. I'm reminding you of the subjectivity of it. However nebulous the definition of humor may be, if I laugh at something then you can be assured that to me it IS funny. And as I laugh heartily, I'm not asking you your opinion of my tastes. If you're not laughing, then evidently it's not funny to you at all. Your laughs are your business, but mine aren't yours. They are my business, and mine alone.

    ReplyDelete
  121. And by the way, you're being a defensive knee-jerk about all this. The point where you think you have to dismiss me (or anyone) as a "fucking moron" makes this into an internet-grade debate. That's not a compliment.

    ReplyDelete
  122. gosh, here i thought calling you a fucking moron was the pinnacle of elegant and civilized discussion, you fucking moron.

    something cannot be subjective by definition unless the definition contains the word 'subjectivity.' you called it subjective by definition. therefore you're telling us what the definition is--and your definition is wrong. now go eat a dick.

    ReplyDelete
  123. No, I didn't define humor. I pointed out that its subjectivity is demonstrated empirically. Surely you're not about to dictate to me when I should laugh? Even if you thought you should, it wouldn't make a lick of difference: I'd continue to laugh when things strike me as funny. You have no more authority in the matter than someone in Tristan da Cunha has over which side of the street Americans should drive on.

    ReplyDelete
  124. This comment has been removed by the author.

    ReplyDelete
  125. you keep assuming that i'm following your definition of humor, and keep assuming that i'm asserting a definition of my own. it's really quite moronic of you.

    congratulations on thinking that being old and still pathetic enough to make the same tired arguments as every other fuckwit fanboy that stumbles on this blog is an accomplishment rather than something to be ashamed of, though!

    ReplyDelete
  126. Once again, I haven't defined humor. Again, it doesn't matter what you think is right or wrong about this. People will laugh at what strikes them as funny. They will not ask you to tell them when to laugh.

    ReplyDelete
  127. go back and read what you've written. you seem to think that 'humor->people laugh at it' and that if people laugh at it it is humor, and if they don't then it is not humor. that's what we in the business call a 'definition,' dipshit.

    ReplyDelete
  128. Plus, I'm not a fanboy. I haven't said a single word for or against xkcd. I rarely read it, and my position in this hasn't a thing to do with a comic strip that bores you in particular.

    ReplyDelete
  129. keep telling yourself that, fanboy

    ReplyDelete
  130. It just doesn't matter if you agree with me or not. People will laugh even if you think they oughtn't.

    ReplyDelete
  131. people will keep murdering even if I think they oughtn't. that doesn't mean I shouldn't denounce them for the monsters they are

    ReplyDelete
  132. People laughing at jokes don't make monsters of themselves for doing so. Are Munroe and his strip stealing the food from your mouth? Are they spreading disease? Have they jinxed your opportunities for advancement at work?

    ReplyDelete
  133. they are most certainly spreading disease of the worst sort.

    ReplyDelete
  134. You poor lad. Thou doth project too much, methinks.

    ReplyDelete
  135. you can't even get your references right. do you have any redeeming qualities at all?

    ReplyDelete
  136. I got my reference just fine. I adapted it to the circumstances at-hand. If you can't see that, then all you're doing at this point is saving face. Have you no dignity?

    Just stop reading the strip. I've been through worse forms of Hell in my life than voluntarily reading drek.

    ReplyDelete
  137. I guess you are probably the type of person who thinks needlessly butchering something is the same thing as adapting it to the circumstances. I keep forgetting you think being old and stupid is an accomplishment.

    ReplyDelete
  138. I'm not old; I'm middle-aged. Of course, to >you< that looks pretty old.

    But really, just stop reading it. My mom and dad both grew up during the Great Depression, when they had to scour the railways for stray lumps of coal just to be able to cook. A civil war has been ongoing in the Congo that's killed as many so far as the Death Camps killed Jews and Slavs. There are and have been people by the countryful who've suffered worse trials than you ever will. Your infatuation with the un-greatness of one comic strip pales.

    ReplyDelete
  139. and you still think it's an accomplishment. not sure why you're having trouble with this. you believe that being old makes you a better person, when, in your case, it obviously has only made you dumber.

    what is with people and thinking that knowing people who lived through some stereotypical tragedy makes complaining about anything else inconsequential? talk about trivializing tragedy. "I know people who were poor and had to scrounge once, therefore your argument is invalid." congratulations, you've reduced your parents' suffering to a useless token in an internet debate.

    if I were an utter tool like yourself I'd be happy to compare my suffering-penis size with you, but, again, not a complete and utter tool. but thanks for playing!

    ReplyDelete
  140. it has an interesting corollary though, if you're losing an argument can you just punch yourself in the face a few times and start winning

    ReplyDelete
  141. it also means the jews automatically win every argument

    ReplyDelete
  142. Whatever you think of me, fact is that xkcd is nothing for you to devote an entire crusade to. It's chicken shit. Ever read Lew Wallace's Ben-Hur? I and a lot of others think it's whole style of prose and story organization is pathetic. And yet whole generations have held it up as a model of high literature. Well, I'm not about to waste my time telling people not to read Ben-Hur. That's none of my business. What other people laugh at is none of yours.

    ReplyDelete
  143. so why are you telling me to stop laughing at xkcd, if it's none of your business what other people like doing?

    ReplyDelete
  144. a CRUSADE! man that makes it sound like we get paid for this shit

    ReplyDelete
  145. I'm not telling you to stop laughing at the strip. I already know that you don't laugh at it these days. What I recommended is that you just stop reading it if it's not to your tastes.

    ReplyDelete
  146. Fuck, I never thought of that!

    Rob why didn't you tell us all that was an option

    ReplyDelete
  147. sorry, maybe I should have clarified. I keep forgetting I'm talking to a functional illiterate. I mean, of course, that you are telling me to stop laughing at how /incredibly fucking stupid/ the strip and its fans are. but you just said it's not your business what people laugh at, so really you're being the worst sort of hypocrite (viz, a stupid one)

    ReplyDelete
  148. "Rob why didn't you tell us all that was an option"

    Of course it's an option. But it appears that Rob refuses to just stop reading, even though he says it's boring. When it was obvious to me that television is a waste of my time, I just stopped watching TV. But I never put up a blog to tell the Constituency that to me it's boring.

    ReplyDelete
  149. "but you just said it's not your business what people laugh at, so really you're being the worst sort of hypocrite"

    Perhaps; but you definitely are one. Whereas I'm not trying to dictate taste to you, your mission here IS to tell people what is and isn't funny. My point tonight is how much you make yourself the very sort of thing you purport to condemn, and you've gotten to the point where you think you can call a boring comic strip a "disease".

    ReplyDelete
  150. Well, tell you what: I'm a couple of thousand miles east of Rob, and it's getting pretty late here, so I gotta go.

    ReplyDelete
  151. you have to specify which part of Rob's gargantuan form you are a couple of thousand miles east of

    CAPTCHA: atrap. It's an atrap, Judy, and you've been CAUGHT!

    ReplyDelete
  152. actually my mission here is to make fun of a comic strip that sucks (key word: 'fun'). yours is to tell people that their form of enjoyment is terrible and blah blah etc.

    I've never personally been opposed to telling people what's good and what isn't. you're the one who's all 'HUMOR IS OBJECTIVELY SUBJECTIVE.' I'm just asking that people follow their own beliefs. notably: if you don't like this blog, don't read it, and if you don't think people should tell other people what they're allowed to find funny, stop telling people what they're allowed to find funny.

    ReplyDelete
  153. enough beeg! i['ll see you all in randall's sweaty mood ring!!

    ReplyDelete