[R] problem with the plm package (2)

Spencer Graves spencer.graves at pdf.com
Thu Nov 30 05:53:23 CET 2006


      I haven't seen any replies to this, so I will offer some 
suggestions even though I'm not familiar with 'plm'. 

      1.  The standard formula syntax in S-Plus and R is that "y ~ x" is 
a request to estimate b0 and b1 in y = b0+b1*x+error.  Since this 
already includes "1", it is essentially equivalent to "y~1+x".  To 
estimate a noconstant model, use something like "y~x-1". 

      2.  Have you tried an email directly to the official "plm" 
maintainer?  An email address for this can be found using 
help(package="plm"). 

      3.  Panel data is a special case of mixed effects data.  The best 
software I know for that the 'nlme' package, which is part of the 
standard R distribution.  The best book on the subject that I know is 
Pinheiro and Bates (2000) Mixed-Effects Models in S and S-Plus 
(Springer).  The "~R\library\nlme\scripts" subdirectory of your R 
installation directory contains files with names like "ch01.R", 
"ch02.R", etc.  These provide R scripts for producing the analyses in 
the book.  You can make local copies of these scripts and work through 
them line by line, trying different things, etc.  The R syntax is 
slightly different from that in the book, e.g., x^2 in a formula must be 
specified as I(x^2).  These differences are overall very minor and easy 
to learn.  However, using exactly what appears in the book without 
modification will in a few cases be misinterpreted, leading to very 
different results.  I had a difficult time understanding what was 
happening until I found this. 

      Hope this helps.
      Spencer Graves

Giangiacomo Bravo wrote:
> Thanks a lot for the quick answerd I received and that helped me to 
> solve my first (simple) problem with the plm package.
>
> Unfortunatley, here's another one:
> I do not know why, but I'm unable to estimate a simple model in the 
> form y ~  1 + x
>
> When I try
>  > zz <- plm(y ~ 1 + x , data=mydata)
> I obtain
> Errore in X.m[, coef.within, drop = F] : numero di dimensioni errato
> which means that the number of dimensions is wrong in X.m[, 
> coef.within, drop = F].
>
> However, I have no problem to estimate a more complex model, e.g.
>  > zz <- plm(y ~ 1 + x + I(x^2), data=mydata)
> in this case everything is ok.
>
> What's wrong?
>
> Thank you,
> Giangiacomo
>
> ______________________________________________
> R-help at stat.math.ethz.ch 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.
>



More information about the R-help mailing list