R help - Apr 2013 - Full Information Maximum Likelihood estimation method for multivariate sample selection problem

Dear R experts/ users

Full Information Maximum Likelihood (FIML) estimation approach is
considered robust over Seemingly Unrelated Regression (SUR) approach
for analysing data of multivariate sample selection problem. The zero
cases in my dependent variables are resulted from three sources:
Irreverent options, not choosing due to negative utility and not used
in the reported time. FIML can address the estimation problem
associated with
cross-equation correlation of the errors.
I am interested to learn and apply the FIML method of estimation. I
searched R resources in internet but I could not get the materials
specific to address my following questions. I request R experts/ users
to address the following queries.

Q.1. hick package of R (e.g. lavaan, mvnmle , stat4 and sem) is
appropriate to analyse the multivariate sample selection problem by
using FIML estimation method?

Q.2. How should it be formulated the code to execute the FIML method ?

Q.3. what is the right method similar to log likelihood ratio to
determine variables stability in the model?

Q.4. My original data of dependent variables are in percentage in
measurement. Do I need to change them any other specific functional
form?

I attempted to formulate the data in the following structure.

Selection equation

ws = c(w1, w2, w3)
# values of dependent variables in selection equations are binary  (1 and 0)

zs = c(z1, z2, z3, z4, z5) # z1, z2, z3 continuous and z4 and z5
dummies explanatory variables in selection equation

Level equation (extent of particular option use)

ys = c(y1, y2, y3)
# values of dependent variables are percentage with some zero cases

xs = c(x1, x2, x3, x4, x5) # x1, x2, x3 continuous and x4 and x5
dummies dependent variables.


Note: The variables in both selection and level equations are mostly same.



Advance thanks for helping me.

Champak Ishram

	[[alternative HTML version deleted]]