Rchoice News

CHANGES IN VERSION 0.3-6

NEW FEATURES

• Adding the Hessian to ivplm.

• Fixing bug with the LR test in "hetprob()"

CHANGES IN VERSION 0.3-5

NEW FEATURES

• New urls in Description.

• New README.

• New methods for effect.

• Improved documentation.

• AIC and BIC for different classes are dropped. We just need "logLik()".

• Changing the behavior of set.seed as a good practice.

CHANGES IN VERSION 0.3-4

NEW FEATURES

• Fixing bug with mtable.

CHANGES IN VERSION 0.3-3

NEW FEATURES

• Add hetprob function to estimate heterokedastic binary models.

• Add ivplm function to estimatate instrumental variables probit model by Maximum Likelihood.

• Fix minor bug in the model.frame.rFormula function with new update of plm.

• Slightly improved documentation.

CHANGES IN VERSION 0.3-2

NEW FEATURES

• Fixing errors and warnings

CHANGES IN VERSION 0.3-1

NEW FEATURES

• Including CITATION

CHANGES IN VERSION 0.3

FIXED BUGS

• fixed minor bug in the model.frame.rFormula function when the id variable is specified using pdata.frame function

NEW FEATURES

• the package not longer includes a vignette. The documentation can be dowloaded at http://msarrias.weebly.com/rchoice-package-in-r.html

• fix some notes regarding to the AIC and BIC methods

• statistic in summary is now labeled as z-value instead of t-value

• cov.Rchoice, cor.Rchoice and se.cov.Rchoice functions are now deprecated. All these functions are now available in the generic function vcov

• slightly improved documentation

CHANGES IN VERSION 0.2

NEW FEATURES

• panel data or longitudinal data is now allow

• added the function se.cov.Rchoice to compute the standard error of the variance-covariance matrix of random parameters

• more flexibility to model hirarchical variables. The argument mvar can be used to choose which variables enter in the mean of the random parameters.

• added the Johnson's Sb function for random parameters.

• added the rFormula and related functions

• initial values for standard deviation of random parameters are set to 0.1. Previously was 0. The argument init.ran can be used to set the initial values. For example init.ran = 0 will set all the initial values for the standard deviation paramters at 0

• slightly improved documentation