Google-Scholar search on AUTHORS on REGRESSION related topics (Part I)

Discussion in 'Scientific Statistics Math' started by Reef Fish, Jul 23, 2006.

  1. Reef Fish

    Reef Fish Guest

    This mini-series of posts on Google-Schoar search on AUTHORS
    was prompted by Richard Ulrich's repeated blame of his MISUSE of
    statistical terms to Jacob Cohen (whom I had never heard of before
    except Rodney Dangerfield who was named Jacob Cohen), and
    THEN made the grossly exaggerated claim:

    RU > Using Google-Scholar with < multiple-regression > or
    RU> <correlation> pulls up books and articles with those in the title
    RU> or abstract. Cohen's textbook on regression and correlation is
    RU > the apparent leader in citations for each of those subjects,
    RU > with 4614.

    while Richard was defending his MISUSE of the term MODE, and
    his error of (1 - type-2 error), the latter blamed on Jacob Cohen,
    when both of those Ulrich errors are NOT regression related.

    That only prompted me to remark:

    RF> Who is "Google-Scholar"? The result you claim is so absurd
    RF> that I can only call it "laughable" but not funny.

    RF> A Google web search on "multiple regression" books gave
    RF> Results 1 - 10 of about 845,000 for "multiple regression" books.
    RF> Hardly any of the references referred to Jacob Cohen's book.


    Then Jerry Dallal tipped me off that Google-Scholar is the option
    in Google search when one clicks "more" (more than the usual
    web or groups or other most-often-used searches).

    So, I went to Google-Scholar, and did a few searches of AUTHORS
    and KEYWORDS, and I quickly found that the "Scholar" was anything
    BUT a scholar!

    It is a bean-counter of the WORST kind.

    A "bean-counter" is the term often used in academia for the Dept
    chairs and Deans who only count the number of papers published
    by a faculty member without any regard to the QUALITY, or even
    the quantity of PAGES. So, a bean would count Wile's 200-page
    paper proving Fermat's Last Theorem the same as a one-page
    "Letter to the Editor" note in a journal. Or a large number of
    short, insignifant papers as if they are worth the same as one
    substantive paper of any length.

    Why id Google-Scholar a bean counter of the WORST kind?
    Because it counts VERY BADLY. I misses some of the most
    significant and important published works by the AUTHOR who
    was specified to search under the specified keyword!

    That of courseis the reason why Richard Ulrich uses this
    bean-counter to make a case for Jacob Cohen, tried to blame
    his own errors as that of "scholar" Cohen.

    A couple of examples of my search will adequately illusrate my
    points above:

    When I specified the Google-Scholar advanced search for the keyward
    "Interactive Data Analysis", with me as the author, I got a count of 3,
    MISSING one of my two books on the subject; my article in the
    Encyclopedia of Statistical Sciences; and HUNDREDS of references
    by other authors and myself in published journal articles with the
    term "interactive data analysis"!

    A count of 3! Some "Scholar".


    When I specified the Google-Scholar advanced search for keywards
    "applied linear models" with my name in AUTHOR, I got a count of "1",
    because I cited the Neter et al book in my SIAM Review of a different
    book!

    Applied Multivariate Data Analysis, Vol. I: Regression and Experimental
    Design.
    RF Ling - SIAM Review, 1992 - JSTOR
    .... REFERENCE [1] J. NETER, W. WASSERMAN, AND MH KUTNER, (1990),Applied
    Linear Models,
    3rd. ed., Ir- win, Boston. ROBERT E LING Clemson University.

    When I replaced the AUTHOR by "John Neter", I got ZERO hit. The same
    for "Michaal Kutner". ZERO hit.

    A "Scholar"? This one is really FUNNY! It couldn't find the
    authors'
    book with a title of "applied linear model" when asked to search the
    AUTHOR with those keywords, but found that book when searching
    for MY publications with that keyword!


    In another part of this series, I shall tabulate the Google-Scholar
    counts
    on various authors I have RECOMMENDED and whose BOOKS and
    JOURNAL ARTICLES I have used in my Regression/Data-Analysis
    courses.

    I'll tabulate the Google-Scholar <sic for Joker> counts for such
    keywords
    as "Multiple Regression", "Regression Diagnostics", "Data Analysis",
    "Interactive Data Analysis", "Linear Models", "Applied Linear Models",
    and "Model Building" for a bean-counter-joker's view of those
    Regression related areas.

    You may do the same on your favorite author/statistician/Quack. :)

    I should add that of those keywords cited above, and the authors
    I checked -- Tukey, Mosteller, Box, Draper, Neter, Kutner, Cook,
    Weisberg, <Jacob Cohen>, and myself <RF Ling>,

    I was the only one with NON-zero counts in ALL of the categories,
    while "Bob Ling" had ALL ZERO counts in ALL of the categories.

    Now I know why Richard Ulrich kept insisting to refer to me as
    Bob Ling, when he should have refered to me as Reef Fish Bob
    (which is the posting name I used, OR my professional name,
    RF Ling, or my full professional name, used in my books and
    journal publications.

    And finally, what about Richard Ulrich? His Google-Scholar
    counts in the 7 categories I searched matched that of "Bob
    Ling" almost perfectly. "Bob Ling" had straight ZEROS,
    while Richard Ulrich had only 6 out of 7 zeros, and a count
    of 2 in "mulltiple regression" in papers in Psychiatry co-authored
    co-authored with 7 and 5 other authors, respectively.

    I had thought Richard Ulrich's insistence on calling me Bob Ling
    was in his reference to the FLAMES in other newsgroups on
    Bob Ling. Now I see a deeper motive -- Richard Ulrich wants
    everyone to know that by his "Google-Scholar" search, one
    can find that Bob Ling has NO PUBLICATIONS at all, while
    Richard Ulrich had 2 (TWO) -- Wow! (Never mind the
    probably inference one can make that he never wrote a single
    word in those two papers with 12 co-authors nearly all of whom
    are MDs and PhDs and he was the lowly MS, probably given
    name-credit by his bosses for cleaning the psychiatric wards. :)

    -- Reef Fish Bob.
     
    Reef Fish, Jul 23, 2006
    #1
    1. Advertisements

  2. On 23 Jul 2006 10:35:36 -0700, "Reef Fish"

    [snip]
    I figured that Reef Fish Bob would know of it, since I had
    explicitly used it in two different discussions with him, in
    January and February. Other people have used it a few times.

    It's the poor-man's version of a citation index. It only
    counts citations found on the Web, which undercounts what
    is published by a factor of 10 or more.

    I noticed that it does seem to find all of my 70+ articles, so
    I do not complain that it misses a lot.

    I think Bob is put-out because it finds just a few citations to
    his own contributions, including his famous one from 1982.
    Jacob Cohen's article on "Multiple regression as a general
    data-analytic tool" from 1968 has 118, so publication date is
    not the only consideration.
    Bob obviously missed the criterion for inclusion, which is not
    "publication" but "citation." I would agree that counting
    publications rather than citations is the *worst* sort of bean-
    counting. Google-scholar is incomplete, but it is not the worst
    kind of count.

    That "one-page letter" would be listed *if* someone
    cited it later on.


    [snip]
    The reader may compare: my own criticisms of Bob are, generally,
    plain attempts to describe what he does in his posts.

    About me: Bob applies his imagination to invent facts,
    then interpretations.
     
    Richard Ulrich, Jul 23, 2006
    #2
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.