Forum: developers

RE: marginalEffect [ Reply ] By: Ott Toomet on 2009-05-09 12:54	[forum:1519]
Agree. However, where should the generic function be? Econmisc was rather meant to be temporary development place... Move it to sampleSelection as well?

RE: marginalEffect [ Reply ]
By: Arne Henningsen on 2009-05-08 22:25

[forum:1516]

I think that it is a good idea to have the generic function ("marginalEffect") and the default methods for objects returned by this function ("coef.marginalEffect", "print.marginalEffect", "vcov.marginalEffect") in the "econMisc" package. However, the methods for objects originating from the "sampleSelection" package ("marginalEffect.probit", "summary.marginalEffect.probit", "print.summary.marginalEffect.probit") should be in the "sampleSelection" package for several reasons. For instance, if a modification in the "sampleSelection" package breaks the "marginalEffect.probit", the problem would probably immediately reported by "R CMD check" and this can be fixed more easily if the method is in the same package as the function that returns the corresponding object. Furthermore, this seems to be standard in R: think of the many print() methods that are in the packages of the corresponding functions and NOT in the package that includes the generic function "print". What do you think?

RE: marginalEffect [ Reply ] By: Ott Toomet on 2009-04-27 03:05	[forum:1461]
Well, moved to econMisc

marginalEffect [ Reply ] By: Ott Toomet on 2009-04-26 04:35	[forum:1457]
marginalEffect: I feel it is a neat and useful function (although not ready for use yet). However, it does not have much to do with sample selection. Move it to econMisc? I also plan to write a method for glm.