R help - Jan 2007 - solving a structural equation model using sem or other package

I am trying to work my way through the book "Singer, JD and Willett, JB,
Applied Longitudinal Data Analysis. Oxford University Press, 2003"  using
R.  I have the SAS code and S-Plus code from the UCLA site (doesn't include
chapter 8 or later problems).  In chapter 8, there is a structural equation/path
model which can be specified for the sem package as follows

S <- cov(al2) #al2 contains the variables alc1, alc2, alc3, and cons
N <- 1122

modelA.ram <- specify.model()
  f1    -> alc1,  NA,  1
  f1    -> alc2,  NA,  1
  f1    -> alc3,  NA,  1
  f2    -> alc1,  NA,  0
  f2    -> alc2,  NA,  .75
  f2    -> alc3,  NA,  1.75
  cons  -> f1,    p0,  1
  cons  -> f2,    p1,  1
  alc1 <-> alc1,  u1,  1
  alc2 <-> alc2,  u2,  1
  alc3 <-> alc3,  u3,  1
  cons <-> cons,  u4,  1
  f1   <-> f1,    s1,  1
  f2   <-> f2,    s2,  1
  f1   <-> f2,    s3,  1 

modelA <- sem(modelA.ram, S, N, analytic.gradient=FALSE)

An equivalent specification in SAS produces the solution presented in the book. 
The variable cons is a constant vector of 1's.  The problem with the sem
package is that the covariance matrix which includes the variable cons is
singular and sem says so and will not continue.  Is there an alternative way to
specify this problem for sem to obtain a solution?  If not, is there another
package that would produce a solution?

Thanks,

Dan Nordlund
Bothell, WA

This is an extract from the sem help page, which deals with your situation:

S covariance matrix among observed variables; may be input as a
symmetric matrix, or as a lower- or upper-triangular matrix. S may
also be a raw (i.e., ``uncorrected'') moment matrix ? that is, a
sum-of-squares-and-products matrix divided by N. This form of input is
useful for fitting models with intercepts, in which case the moment
matrix should include the mean square and cross-products for a unit
variable all of whose entries are 1; of course, the raw mean square
for the unit variable is 1. Raw-moment matrices may be computed by
raw.moments.

On 24/01/07, Daniel Nordlund <res90sx5 at verizon.net>
wrote:
> I am trying to work my way through the book "Singer, JD and Willett,
JB, Applied Longitudinal Data Analysis. Oxford University Press, 2003" 
using R.  I have the SAS code and S-Plus code from the UCLA site (doesn't
include chapter 8 or later problems).  In chapter 8, there is a structural
equation/path model which can be specified for the sem package as follows
>
> S <- cov(al2) #al2 contains the variables alc1, alc2, alc3, and cons
> N <- 1122
>
> modelA.ram <- specify.model()
>   f1    -> alc1,  NA,  1
>   f1    -> alc2,  NA,  1
>   f1    -> alc3,  NA,  1
>   f2    -> alc1,  NA,  0
>   f2    -> alc2,  NA,  .75
>   f2    -> alc3,  NA,  1.75
>   cons  -> f1,    p0,  1
>   cons  -> f2,    p1,  1
>   alc1 <-> alc1,  u1,  1
>   alc2 <-> alc2,  u2,  1
>   alc3 <-> alc3,  u3,  1
>   cons <-> cons,  u4,  1
>   f1   <-> f1,    s1,  1
>   f2   <-> f2,    s2,  1
>   f1   <-> f2,    s3,  1
>
> modelA <- sem(modelA.ram, S, N, analytic.gradient=FALSE)
>
> An equivalent specification in SAS produces the solution presented in the
book.  The variable cons is a constant vector of 1's.  The problem with the
sem package is that the covariance matrix which includes the variable cons is
singular and sem says so and will not continue.  Is there an alternative way to
specify this problem for sem to obtain a solution?  If not, is there another
package that would produce a solution?
>
> Thanks,
>
> Dan Nordlund
> Bothell, WA
>
> ______________________________________________
> 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.
>

-- 
================================David Barron
Said Business School
University of Oxford
Park End Street
Oxford OX1 1HP