[R] effect function in the effects package

[John Fox]

Dear Xu Jun,

On Thu, 20 Oct 2011 14:41:30 -0400
 Xu Jun <junxu.r at gmail.com> wrote:
> Dear Professor Fox,
> 
> Now I got it. It all comes from my unfamiliarity with the effect
> function. I forgot the c part in the given.values option, plus it
> looks like plot(effect()) does not like factor(warm) in the polr
> function. So here are the two working lines:
> 
> ordwarm2$warm2 <- as.factor(ordwarm2$warm)

Without the data, I have no idea what's going on here; the implication is that warm wasn't a factor.

> myologit <- polr(warm2 ~ yr89 + male + white + age + ed + prst,
>               data=ordwarm2, method=c("logistic"))
> plot(effect("age", myologit, xlevels=list(age=seq(20, 80, 5)),
> given.values=(c(male=1, yr89=1))))

given.values is an argument. c() is a standard R function for combining values into a vector (or a list); it is not particular to the effects package. The parentheses around (c(male=1, yr89=1)) are entirely unnecessary, and the call to c() in  c("logistic") does nothing.

Best,
 John

> 
> Again, thanks a lot for your effects package that makes graphing so much easier.
> 
> Jun
> 
> On Thu, Oct 20, 2011 at 11:55 AM, John Fox <jfox at mcmaster.ca> wrote:
> > Dear Xu Jun,
> >
> > I'm not sure whether this is the source of the error, but it may help to spell the xlevels argument correctly (it is not "xlevles").
> >
> > I hope this helps,
> >  John
> >
> > ------------------------------------------------
> > John Fox
> > Sen. William McMaster Prof. of Social Statistics
> > Department of Sociology
> > McMaster University
> > Hamilton, Ontario, Canada
> > http://socserv.mcmaster.ca/jfox/
> >
> > On Thu, 20 Oct 2011 10:34:30 -0400
> >  Xu Jun <junxu.r at gmail.com> wrote:
> >> Dear r-help listers,
> >>
> >> I am using effects to produce an effect plot after the proportional
> >> odds logistic regression model. There is no problem for me to estimate
> >> the model, but when it comes to the graphing, I was stuck. see the
> >> codes below:
> >>
> >> ##############################################################################
> >>  myologit <- polr(factor(warm) ~ yr89 + male + white + age + ed + prst,
> >> +               data=ordwarm2, method=c("logistic"))
> >>
> >> require(effects)
> >> plot(effect("age", myologit, xlevles=list(age=seq(20, 80, 5),
> >> given.values(male=1, yr89=1))))
> >>
> >> ##############################################################################
> >>
> >> I got the following error message:
> >>
> >> Warning message:
> >> package 'effects' was built under R version 2.13.2
> >> Error in plot(effect("age", myologit, xlevles = list(age = seq(20, 80,  :
> >>   error in evaluating the argument 'x' in selecting a method for
> >> function 'plot': Error in model.frame.default(formula = factor(warm) ~
> >> yr89 + male + white +  :
> >>   variable lengths differ (found for 'yr89')
> >>
> >>
> >> the full set of codes:
> >>
> >> > myologit <- polr(factor(warm) ~ yr89 + male + white + age + ed + prst,
> >> +               data=ordwarm2, method=c("logistic"))
> >>
> >> > summary(myologit)
> >>
> >> Re-fitting to get Hessian
> >>
> >> Call:
> >> polr(formula = factor(warm) ~ yr89 + male + white + age + ed +
> >>     prst, data = ordwarm2, method = c("logistic"))
> >>
> >> Coefficients:
> >>           Value Std. Error t value
> >> yr89   0.523912   0.079899   6.557
> >> male  -0.733309   0.078483  -9.344
> >> white -0.391140   0.118381  -3.304
> >> age   -0.021666   0.002469  -8.777
> >> ed     0.067176   0.015975   4.205
> >> prst   0.006072   0.003293   1.844
> >>
> >> Intercepts:
> >>     Value    Std. Error t value
> >> 1|2  -2.4654   0.2389   -10.3188
> >> 2|3  -0.6309   0.2333    -2.7042
> >> 3|4   1.2618   0.2340     5.3919
> >>
> >> Residual Deviance: 5689.825
> >> AIC: 5707.825
> >>
> >> > plot(effect("age", myologit, xlevles=list(age=seq(20, 80, 5), given.values(male=1, yr89=1))))
> >> Warning message:
> >> package 'effects' was built under R version 2.13.2
> >> Error in plot(effect("age", myologit, xlevles = list(age = seq(20, 80,  :
> >>   error in evaluating the argument 'x' in selecting a method for
> >> function 'plot': Error in model.frame.default(formula = factor(warm) ~
> >> yr89 + male + white +  :
> >>   variable lengths differ (found for 'yr89')
> >>
> >> ______________________________________________
> >> R-help at 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.
> >
> >
> >
> >

------------------------------------------------
John Fox
Sen. William McMaster Prof. of Social Statistics
Department of Sociology
McMaster University
Hamilton, Ontario, Canada
http://socserv.mcmaster.ca/jfox/