Re: [R] Extract model matrix from plm

There is actually a model.matrix method for plm objects. Unfortunately, it is not currently exported and therefore can't be used by the end user of plm. I'll fix that in the next update. However, you won't find in this matrix the dummies because demeaned values are used. If you really want the

Re: [R] Panel data with binary dependent variable

To my knowledge, fixed and random effect models may be estimated for the logit model and only the random effect model for the probit model (because of the incidental parameter problem). I think clogit in the survival package fits the model that is called the fixed effect logit model in the

Re: [R] Truncated Normal Distribution and Truncated Pareto distribution

The truncreg package fits the truncated normal model. Le lundi 19 avril 2010 à 00:21 -0700, Julia Cains a écrit : Dear R helpers, I have a bimodal dataset dealing with loss amounts. I have divided this dataset into two with the bounds for the first dataset i.e. dataset-A being 5,000$ to

Re: [R] plm package twoways effect problem

It's difficult to guess what the pb is without the data. Try to estimate a 'simple' model, ie without the dummies (something like : dsnfi3LP=plm(lnLP~lnC+lnL+lnM +lnE,data=newdata,effect=twoways,model=within) Try also pdim(newdata) to check if you don't have 1 observation for some individuals.

Re: [R] mlogit

In the code of the mlogit function, the call is saved under the name 'mldata' that you use to store the result of the mlogit.data function. When calling mlogit with 'mldata' as the data argument, there is a conflict between these two objects with the same name, one of class 'call' and one of class