I'm wondering if there is a simple way to use Zelig to compare changes in
the differences between two groups across conditions. I have two groups, I
can easily determine that the difference between them in condition1 is
signfiicant and also that the difference between them in condition2 is
significant. But I am wondering wether the change in the difference is
signfiicant -- that is, does the change from condition1 to condition2 make
them significantly more or less alike?
Donald Braman
http://www.law.gwu.edu/Faculty/profile.aspx?id=10123http://research.yale.edu/culturalcognitionhttp://ssrn.com/author=286206
Dear Amber,
Please install the latest version of Zelig by starting R and typing:
install.packages("Zelig", repos = "http://gking.harvard.edu/")
The version of Zelig that you are using is quite old -- and hence,
survey models are not available.
Best,
Olivia
On Tue, Jun 24, 2008 at 12:59 PM, Amber Wichowsky <wichowsky(a)wisc.edu> wrote:
> Hello,
> I have updated Zelig on my home computer (Mac) and work computer (PC). I am
> able to run survey.logit on my Mac, but I get the following message on the
> PC:
>
> Error in zelig.default(PVOTE1 ~ ZNSLATEN + ZFP2 + ZWWLATEN + ZRACE2 + :
> logit.survey not supported. Type help.zelig("models") to list
> supported models.
>
> when I type help.zelig("models"), the survey.logit and survey.normal models
> are not listed. I am running the following version of Zelig:
>
> Zelig (Version 3.0-6, built: 2007-10-03)
>
> It is running on R 2.5.1
>
> Thanks very much,
> Amber Wichowsky
>
> Graduate Student
> Department of Political Science
> University of Wisconsin-Madison
>
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
Hello,
I have updated Zelig on my home computer (Mac) and work computer (PC). I am
able to run survey.logit on my Mac, but I get the following message on the
PC:
Error in zelig.default(PVOTE1 ~ ZNSLATEN + ZFP2 + ZWWLATEN + ZRACE2 + :
logit.survey not supported. Type help.zelig("models") to list
supported models.
when I type help.zelig("models"), the survey.logit and survey.normal models
are not listed. I am running the following version of Zelig:
Zelig (Version 3.0-6, built: 2007-10-03)
It is running on R 2.5.1
Thanks very much,
Amber Wichowsky
Graduate Student
Department of Political Science
University of Wisconsin-Madison
I ran a least squares model with Zelig and received results. However,
the last variable I listed in the model was dropped, and I received
the following waring messages.
1: In cbind(coef1, tmp[, 1]) :
number of rows of result is not a multiple of vector length (arg 2)
2: In cbind(se1, tmp[, 2]) :
number of rows of result is not a multiple of vector length (arg 2)
If I eliminate the offending variable from the model the errors no
longer appear.
Thoughts? Thanks much.
Best,
-c
--
Casey A. Klofstad
University of Miami
Department of Political Science
Coral Gables, FL
klofstad(a)gmail.com
http://moya.bus.miami.edu/~cklofstad
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
I am running a logistic regression combining five imputed datasets using
Zelig. I just noticed small but nontrivial differences between the model
output using Zelig (Version 2.8-3, built: 2007-05-29) and Zelig (Version
3.0-1, built: 2007-07-24).
I ran logistic regressions on each of the individual imputed datasets in
spss and compared it to the individual outputs for each imputed dataset that
Zelig provides. Version 3.0-1 closely matches the spss regressions, while
version 2.8-3 is a bit off.
In what ways do these two versions differ?
Thanks very much,
Amber Wichowsky
Ph.D Student
Department of Political Science
University of Wisconsin-Madison
Hi all:
I'm having some trouble estimating a seemingly unrelated regression
(sur) model in Zelig. I get the same error message when I run my model
with my data *and* when I run the model in the Zelig help files (p523 of
the manual), which suggests it might be a more general problem than just
my implementation.
Here's the code I ran (from the help files):
data(grunfeld)
formula <- list(mu1 = Ige ~ Fge + Cge, mu2 = Iw ~ Fw + Cw)
z.out <- zelig(formula = formula, model = "sur", data = grunfeld)
And here's the error message I get:
Error in systemfit(data = data, eqns = formula, method = method, inst =
inst, :
argument "formula" is missing, with no default
I don't think it's a naming problem -- the error persists even if I
specify that "eqns=formula", for example. I have the latest version of
Zelig, too. Does anyone have any idea what might be going wrong?
Thanks, Phil
--
Ph.D. Candidate
Harvard University Department of Government
1737 Cambridge Street, Cambridge, MA 02138
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
Dear Keith,
I fixed the problem that you identify in plot.ci (a simple if
statement problem), but there's another associated problem in setx
that prevents plot.ci from working with mixed effect models (for now).
As a sidebar, your syntax below doesn't work for plot.ci in general.
Here's why:
What you have below compares expected values for just one change in an
independent variable. These would look like just two bars on the plot
showing the range of the 95% confidence intervals. A more informative
way to present this information would be to do two overlapping density
plots.
To produce this plot, run demo(logit.mixed), then:
s.high <- sim(z.out, x = x.high)
s.low <- sim(z.out, x = x.low)
xlim <- range(c(s.high$qi$ev, s.low$qi$ev))
plot(density(s.high$qi$ev), col = "red", xlim = xlim)
lines(density(s.low$qi$ev), col = "blue")
In contrast, the plot.ci function is intended to show the change in
the dependent variable for more than one change in an independent
variable. For example, if you let age or education vary from 10:20,
you could use plot.ci rather than the procedure above. See
demo(plot.ci) for this example.
Yours,
Olviia
On Wed, Jun 11, 2008 at 9:04 PM, Keith Schnakenberg
<keith.schnakenberg(a)gmail.com> wrote:
> Hello all,
>
> I noticed that Olivia posted in January that she added mixed effects models
> to plot.ci and that it would be in the next release. Has it been added? I am
> having trouble using it on the latest version. A reproducible example is
> given below:
>
>> data(voteincome)
>> x.high <- setx(z.out1, education = quantile(voteincome$education, 0.8))
>> x.low <- setx(z.out1, education = quantile(voteincome$education, 0.2))
>> s.out1 <- sim(z.out1, x = x.high, x1 = x.low)
>> plot.ci(s.out1, CI=95, qi="ev")
> Error in plot.ci(s.out1, CI = 95, qi = "ev") :
> plot.ci() is valid only for non-categorical, univariate response models.
>
> Thanks,
> Keith Schnakenberg
>
> -
> Zelig Mailing List, served by Harvard-MIT Data Center
> Send messages: zelig(a)lists.gking.harvard.edu
> [un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
> Zelig program information: http://gking.harvard.edu/zelig/
>
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
Hello all,
I noticed that Olivia posted in January that she added mixed effects
models to plot.ci and that it would be in the next release. Has it
been added? I am having trouble using it on the latest version. A
reproducible example is given below:
> data(voteincome)
> x.high <- setx(z.out1, education = quantile(voteincome$education,
0.8))
> x.low <- setx(z.out1, education = quantile(voteincome$education,
0.2))
> s.out1 <- sim(z.out1, x = x.high, x1 = x.low)
> plot.ci(s.out1, CI=95, qi="ev")
Error in plot.ci(s.out1, CI = 95, qi = "ev") :
plot.ci() is valid only for non-categorical, univariate response
models.
Thanks,
Keith Schnakenberg
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
