Grokbase Groups R r-help March 2002
FAQ
Hello everybody,

a question that connect to the question of Frederik Karlsons about 'how
to stand. betas'
With the stand. betas i can compare the influence of the different
explaning variables. What do i with the betas of factors? I can't use
the solution of JohnFox, because there is no sd of an factor. How can i
compare the influence of the factor with the influence of the numeric
variables?

Thanks for any help

joerg

--
Joerg Maeder .:|:||:..:.||.:: maeder at atmos.umnw.ethz.ch
Tel: +41 1 633 36 25 .:|:||:..:.||.::
http://www.iac.ethz.ch/staff/maeder
PhD student at INSTITUTE FOR ATMOSPHERIC AND CLIMATE SCIENCE (IACETH)
ETH Z?RICH Switzerland
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Search Discussions

  • John Fox at Mar 5, 2002 at 1:37 pm
    Dear Joerg,
    At 09:40 AM 3/5/2002 +0100, you wrote:

    a question that connect to the question of Frederik Karlsons about 'how
    to stand. betas'
    With the stand. betas i can compare the influence of the different
    explaning variables. What do i with the betas of factors? I can't use
    the solution of JohnFox, because there is no sd of an factor. How can i
    compare the influence of the factor with the influence of the numeric
    variables?
    The simple function in my response to Frederik Karlsons's question will
    work with a model that includes factors, producing a standardized
    coefficient for each column of the model matrix (except the constant), but
    the result is not sensible, depending for example on the contrast coding. I
    don't believe that there is a sensible answer to your question, and
    moreover, I would argue that standardized coefficients are not even
    terribly useful for comparing the "influence" of quantitative predictors.

    By the way, the function that I included in my previous message
    inadvertently used the response function in my car library to extract the
    response variable from the model. Without this library attached, one could
    substitute model.response(model.frame((mod))).

    Regards,
    John
    -----------------------------------------------------
    John Fox
    Department of Sociology
    McMaster University
    Hamilton, Ontario, Canada L8S 4M4
    email: jfox at mcmaster.ca
    phone: 905-525-9140x23604
    web: www.socsci.mcmaster.ca/jfox
    -----------------------------------------------------

    -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
    r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
    Send "info", "help", or "[un]subscribe"
    (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
    _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
  • Vito muggeo at Mar 5, 2002 at 1:49 pm

    a question that connect to the question of Frederik Karlsons >about 'how
    to stand. betas'
    With the stand. betas i can compare the influence of the >different
    explaning variables. What do i with the betas of factors? I can't >use
    the solution of JohnFox, because there is no sd of an factor. >How can i
    compare the influence of the factor with the influence of the >numeric
    variables?

    Use the t-values from summary(lm(....))!
    Are not the standardized regression coefficients the "t values"?

    cheers,
    vito


    ----- Original Message -----
    From: "Joerg Maeder" <maeder@atmos.umnw.ethz.ch>
    To: "'R-help at lists.R-project.org'" <r-help@stat.math.ethz.ch>
    Sent: Tuesday, March 05, 2002 9:40 AM
    Subject: [R] enhanced Question to stand. Beta


    Hello everybody,

    a question that connect to the question of Frederik Karlsons about 'how
    to stand. betas'
    With the stand. betas i can compare the influence of the different
    explaning variables. What do i with the betas of factors? I can't use
    the solution of JohnFox, because there is no sd of an factor. How can i
    compare the influence of the factor with the influence of the numeric
    variables?

    Thanks for any help

    joerg

    --
    Joerg Maeder .:|:||:..:.||.:: maeder at atmos.umnw.ethz.ch
    Tel: +41 1 633 36 25 .:|:||:..:.||.::
    http://www.iac.ethz.ch/staff/maeder
    PhD student at INSTITUTE FOR ATMOSPHERIC AND CLIMATE SCIENCE (IACETH)
    ETH Z?RICH Switzerland
    -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.
    -.-
    r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
    Send "info", "help", or "[un]subscribe"
    (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
    _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._.
    _._
    -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
    r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
    Send "info", "help", or "[un]subscribe"
    (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
    _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
  • Thomas Lumley at Mar 5, 2002 at 4:21 pm

    On Tue, 5 Mar 2002, vito muggeo wrote:

    a question that connect to the question of Frederik Karlsons >about 'how
    to stand. betas'
    With the stand. betas i can compare the influence of the >different
    explaning variables. What do i with the betas of factors? I can't >use
    the solution of JohnFox, because there is no sd of an factor. >How can i
    compare the influence of the factor with the influence of the >numeric
    variables?

    Use the t-values from summary(lm(....))!
    Are not the standardized regression coefficients the "t values"?
    Not at all. The t-values are coefficients divided by their standard
    errors. Standardised coefficients are divided by the standard deviation of
    the predictor.

    There isn't really a way to do this for factors.

    Consider the much simpler case of binary data. You could use the standard
    deviation of the variable, but there might be just as strong a case for
    considering the variable as a discretisation of a Normal and working out
    some sort of latent variable model. These approaches would typically give
    quite different results, suggesting that the question doesn't have a
    unique answer.


    -thomas



    -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
    r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
    Send "info", "help", or "[un]subscribe"
    (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
    _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
  • John Fox at Mar 5, 2002 at 4:27 pm
    Dear Vito,
    At 02:49 PM 3/5/2002 +0100, vito muggeo wrote:
    a question that connect to the question of Frederik Karlsons >about 'how
    to stand. betas'
    With the stand. betas i can compare the influence of the >different
    explaning variables. What do i with the betas of factors? I can't >use
    the solution of JohnFox, because there is no sd of an factor. >How can i
    compare the influence of the factor with the influence of the >numeric
    variables?

    Use the t-values from summary(lm(....))!
    Are not the standardized regression coefficients the "t values"?
    No: The t-values are the ratios of the coefficients to their standard
    errors. What's usually meant by "standardized coefficients" are the
    coefficients obtained when the variables (both response and predictors) are
    standardized to unit variance.

    John

    -----------------------------------------------------
    John Fox
    Department of Sociology
    McMaster University
    Hamilton, Ontario, Canada L8S 4M4
    email: jfox at mcmaster.ca
    phone: 905-525-9140x23604
    web: www.socsci.mcmaster.ca/jfox
    -----------------------------------------------------

    -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
    r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
    Send "info", "help", or "[un]subscribe"
    (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch
    _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupr-help @
categoriesr
postedMar 5, '02 at 8:40a
activeMar 5, '02 at 4:27p
posts5
users4
websiter-project.org
irc#r

People

Translate

site design / logo © 2017 Grokbase