On 16/01/2009, at 1:50 AM, gregor rolshausen wrote:
hello.
I have a question on the interpretation of a logistic model.
is it helpful to exponentiate the coefficients (estimates)? I think I
once read something about that, but I cannot remember where.
if so, how would be the interpretation of the exp(estimate) ?
exp(beta_i) is the odds ratio for success when the i-th predictor
x_i is incremented by 1.
In particular if x_i is a 0-1 indicator variable then exp(beta_i) is
the odds ratio for comparing the odds of success when x_i = 1 with
the odds of success when x_i = 0.
E.g. if x_i = 0 for Male and x_i = 1 for Female, and exp(beta_i) = 2,
then the odds of success for Females are twice as great as the odds
of success for Males. I.e. Females are ``twice as likely'' to succeed
as Males, all other things being equal. (Which may or may not be a
Good
Thing, depending on what ``success'' really means. :-) )
would there be a change of the interpretation of the ANOVA table
(or is
the ANOVA table not really helpful at all?).
ANOVA tables are ***so*** 20th century!
cheers,
Rolf Turner
######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
