Skip to content
 

What Zombies see in Scatterplots

This video caught my interest – news video clip
(from this
post2)

http://www.stat.columbia.edu/~cook/movabletype/archives/2011/02/on_summarizing.html

The news commentator did seem to be trying to point out what a couple of states had to say about the claimed relationship – almost on their own.

Some methods have been worked out for zombies to do just this!

So I grabbed the data as close as I quickly could, modified the code slightly and here’s the zombie veiw of it.

PoliticInt.pdf

North Carolina is the bolded red curve, Idaho the bolded green curve.
Missisipi and New York are the bolded blue.

As ugly as it is this is the Bayasian marginal picture – exactly (given MCMC errror).

K?
p.s. you will get a very confusing picture if you forget to centre the x (i.e. see chapter 4 of Gelman and Hill book)

17 Comments

  1. bxg says:

    What "x"? :-)

    But the fact there seems to be no "x" makes me see … after too long … that there's an intentional joke in your posts here I'm too ignorant to see. Help me anyone please: what type of person, what knowledge does it require, what is the trick if any?, to see the purpose/humor/??? in K? posts.
    I don't have the sophistication needed, and I can no longer attribute it to a deliberately and abnormally opaque style, so I'm just stupid – I think… Statisticians only? Geniuses only? Some other club? Who gets these posts?

  2. Phil says:

    bxg, I can never figure out what he's saying either. But I don't think it's deliberate, I think he's just a really smart guy who doesn't write clearly.

  3. anon says:

    I agree with bxg; this post is impenetrable throughout;

    This video caught my interest – news video clip
    Why? There isn't a scatterplot in the video until it's 1/2 way through. And no zombies at all.

    The news commentator did seem to be trying to point out what a couple of states had to say about the claimed relationship – almost on their own.
    What did they say? About what relationship? And was it the news commentator saying something on their own, or the states, or the states' data on their own?
    Some methods have been worked out for zombies to do just this!
    This idea (of drawing log priors, likelihoods and posteriors) gets shoehorned into your posts wherever possible. Is it really so universally important? Its relevance to the issue here is not obvious – you're commenting on understanding a scatterplot, and scatterplots involves no likelihood, and no prior.

    So I grabbed the data as close as I quickly could, modified the code slightly and here's the zombie veiw of it.
    What data? From where? What code are you modifying? Would it kill you to read through the sentence for some basic grammar and spelling?
    North Carolina is the bolded red curve, Idaho the bolded green curve.
    Missisipi and New York are the bolded blue.

    Two bolded blue curves, three no-bolded blues, various other line types (what does dashed mean?) and no legend. No indication of what beta is, and no indication of any assumed model. For a blog exemplifying statistical graphics, this is just a joke.
    As ugly as it is this is the Bayasian marginal picture – exactly (given MCMC errror).
    Monte Carlo error? Or errors in your MCMC? What is a "Bayesian marginal picture"? But I agree it's an ugly picture, at least.
    p.s. you will get a very confusing picture if you forget to centre the x (i.e. see chapter 4 of Gelman and Hill book)
    You don't tell us what 'x' is, anywhere, so this comment makes no sense at all. Also, getting a confusing picture can not be said "in other words" to be looking up chapter 4 of Gelman and Hill.

    Excepting those from Keith, this blog's postings are of a very high standard. Unless it's an act of sympathy, or a cruel joke, I don't see that posting Keith's streams of consciousness is helping anyone.

    Keith: whatever you write, for any audience, please review it, trying to put yourself in the mind of a reader (a reader who is not Keith).

  4. Andrea says:

    A zombie post deserves one and only one reply: braaaaaaaaaains!

  5. K? O'Rourke says:

    Thanks Phil, but here too much haste and I was relying too much on familiarity of the previous posts that were referred to.

    Also, currently unwilling to reprioritize to take the time and effort to make my blog posts easy to read for a wide audience – hence locating them in the zombie category.

    Here it was post rather than just a comment as thats the only way I am currently aware of how to upload a pdf file that was needed to _show_ my point (which was simply an add on to Andrew's pointing out formulas for individual point slopes in his post)

    But there is R code in my earlier post referred that does work and would allow anyone familiar with R to pin down exactly what I am doing if not also why.

    K?

  6. K? O'Rourke says:

    bxg: Here likely because the blog post was intended or only appropriate for those who already grasped the plots _and_ previous posts about it.

    The "x" though, is pretty standard for refering to the independent variable in regression.

    I am working on material for a wider audience (even an ASA CE proposal) but that nots ready yet – its a lot of work. Those with little background in likelihood (not much more than maximum likelihood estimates and standard errors) find the plots very puzzling, those with extensive knowledge of likelihood find them somewhat trivial or why would you need to see that in a plot – its obvious mathematically.

    The answer to your question "Who gets these posts?" would be helpful to me ~ that sort of question is hard to get a good sense of.

    K?

  7. Alex Cook says:

    K?: you say you don't have the inclination to make your posts readily understandable, but I suspect most readers are similarly disinclined to invest time trying to understand them! (Especially if the links that provide the background don't work…) If you make it easy for the reader to understand, we'll listen to what you have to say; if you don't, we go blur, and the time you spent jotting down your thoughts is wasted.

  8. K? O'Rourke says:

    anon: You are simply epecting more than I willing to put into these posts.

    This has been discussed before and Andrew is confortable with being the one who will remove material that should not remain.

    Years ago I stopped posting to sci.stat.math (or something like that) because of _demands_ for clarification or outright claims of _your totally wrong_.

    Unless you find a bug in my R code or a fact that requires correction, I believe you should just disregard the post.

    With regard to "scatterplots involves no likelihood, and no prior" I believe it does, its just implicit.

    K?

  9. bxg says:

    K?..

    It's awfully hard to even hint at criticism against someone brave and intelligent who posts their thoughts so publically (which makes you/them a better person than I!)

    But, emboldened by earlier responses, I will go there anyway. Your responses to this thread are, in a word, arrogant. Maybe justifiably so, I don't know. Let's just look at one concern, that you can't spare the time to make your "blog posts easy to read for a wide audience". I fear no real risk of contradiction by observing:
    -your posts are utterly not just hard to read, but utterly incomphrensible to a wide audience.
    -and a medium audience
    -and a narrow audience
    … but maybe there's a super-super select small group who gets it (and maybe, but I'd be astonished if || > 5) who furthermore finds it _easy_ to read!

    Yet you give the impression that every second you might (but would not) spend on, say, proofreading, or condescending in the slightest to your audience, is an irrational inconvenience – who knows not what glories of your thought might therefore be lost.

    You write many such posts. Can you not make some comment on WHY you do so? And, in your minds-eye, for WHO? Give us a filter please: let us know that if we are not such-and-such we should please move on. Ideally for some notion of such-and-such that include more than yourself and your students? The nutshell request here is for you step back from details and give us a more philosophic view of why you post, who you want to reach, what type of feedback on your posts you would view as constructive (obviously not anyone's in this thread so far!) & etc.

    Professor Gelman, I do not expect this comment to survive your moderation and perhaps do not think it should. I may make no demands on your (of course not!) but I would humbly request you either reject this post of mine or make some sort of comment on what/how your readers should sensibly approach K? posts. You generally exemplify clear writing (and have spent various posts explicitly around how seriously you take this and how hard you work at it); to let K? share your forum without comment/clarification is … confusing.

  10. bxg says:

    Anon's critique of your post was great – IMO careful and detail-oriented, though perhaps one could take issue with the tone. I suspect he spent more time on his analysis than on your original post. But whatever, it's just clear that he was trying to be helpful and constructive. You may not care at all (later posts suggest you don't) but can't you respect the intent?

    Yes, I concede, scattered among his more many substantive, thoughtful, is the less constructive/most nit-picky gripe: "Would it kill you to read through the sentence for some basic grammar and spelling?"

    But looking at his/her post overall, to respond as you do – adding a (presumably deliberate?) extra large dose of mispellings and grammatical sloppiness – seems just childish.

  11. anon says:

    But there is R code in my earlier post referred that does work and would allow anyone familiar with R to pin down exactly what I am doing if not also why.

    It doesn't allow what you claim. I'm highly familiar with R, and likelihood, and I can't follow it despite trying.

    You don't document what the code is doing, your variable names are no help, you give no link to a concise write-up of what analysis is being done (you link to an entire PhD thesis without even a page reference) and you write loops in the most obtuse way I've ever seen. You can't even be bothered to put a legend on a plot. Like your blog posts, the code seems written for an audience of just one person – you.

    I know I don't have to read your posts – and often I don't. But wouldn't you be better served by explaining your ideas in a way that engages the audience, rather than alienating them?

  12. Phil says:

    There's no point writing a post if nobody can understand it.

  13. Nick Cox says:

    "You are simply epecting more than I willing to put into these posts."

    Evidently!

    I think this is sad or puzzling rather than something to beat up Keith about. He has said some things that I understood and thought interesting, but mostly I too don't know what on earth he is on about, even in the very broadest sense, and I am unwilling to decode the poor presentations. Sometimes it seems as if Keith doesn't read his own postings _once_ for errors!

    Keith: If you put twice as much effort into half as many posts, you would have much more impact.

    Otherwise most people will just ignore or delete your contributions, and flak like this will be repeated.

  14. AC says:

    It's fine to address a post to a limited audience, but I am confused about who Keith's intended audience is. Did anyone understand this post? Keith sort of suggests that this post is just a placeholder for him to store a figure. But a blog isn't supposed to be private notes to oneself. There's supposed to be some audience, and there's no point in uploading to the web in a public forum if you don't want other people to understand what you did.

    Keith argues that this post assumes some prerequisites. But even if one has read Keith's previous zombie course post, the figure in the current post is still unintelligible. Even the figure in the earlier post is unexplained and unintelligible at first glance; deciphering it is evidently a homework exercise. The figure in this latest post isn't even labeled. Sure, Keith points out what a few of the bold curves are, but what about the rest of them?

    Unintelligible figure aside, the post also provides no interpretation. What is the point of the post? What message about union membership and budget deficits are we supposed to take away from Keith's figure that wasn't already apparent in the original scatterplot?

    I'm not really trying to pile on Keith here, but I'm sincerely wondering what the goal of his posts are. I honestly don't know who was supposed to understand and appreciate the current post.

  15. bxg says:

    I suppose I've driven the "flak" in this thread and so if there's "beating up" going on, I suppose it's by me – ?
    In general, I have no problems whatsoever just entirely ignoring blog posts that I don't understand or disagree with or whatever. But this is not the situation here. However may I have come across as, especially in my immediately prior post, I truly don't intend any "beating up" (and I think as an anonymous commenter this would be both contemptuous and presumptuous.) I _sincerely_ would like to know what I would need to do to learn how to decode K?'s posts. And if this is impractical, I truly, without the slightest sarcasm or implicit criticism of anyone, would feel usefully educated in learning a bit about who the actual audience is.

  16. K? O'Rourke says:

    Believe the broken link has been fixed.
    (I do recall checking it before posting.)

    This has become very awkward and Andrew has been very patient and supportive.

    I had informed him that I likely _should_ stop posting and commenting here last Friday – but tried to _sneak_ one last one in.

    I believe I have learned a lot and I certainly don't think any of the comments were out of line – some perhaps a little bit impatient but that often just reflects real interest.

    And of course one last apparently arrogant comment – confusing posts could be embarrassed (ok this typo is just too good to fix) as an opportunity to learn!!

    K?

  17. Phil says:

    On the one hand, I do think that posts like this one are not worth reading and therefore not worth writing. None of us — and I do mean none — can figure out what you're saying.

    On the other hand, I think the problem is with the writing, not the content. I'm still hoping that you'll decide to put more effort into writing your posts, rather than deciding not to do them at all. Don't take your ball and go home; explain the rules of the game so we can all play with your ball.