[R] Help with nlme--freq weights, logit model, and more
SamL
saml at demog.berkeley.edu
Fri Apr 18 18:16:59 CEST 2003
Below you will find the output from a failed multi-level model run. I am
trying to estimate the following model:
Pr(PLFP=1)= logistic regression ->
B1_j * bm + B2_j * wm + B3_j bf + B4_j wf + B5 yrsed+ B6 age+ B7 age^2+e_ij
B1_j = G01 + G11 bmxd + d1
B2_j = G02 + G12 wmxd + d2
B3_j = G03 + G13 bfxd + d3
B4_j = G04 + G14 wfxd + d4
d1-d4 freely correlated
Note that there is no intercept, so the B1-B4 are race/sex specific
intercepts (bm=black male, wm=white male, bf=black female, wf=white
female).
This model allows the xd variables to influence the means within the
geographic locations (50 states).
SO, my questions:
1)I cannot figure out what the error code means. This is the first time I
have received that code, which I think is a good sign--I am getting closer
to the correct syntax. But, still, no results. I am not sure what to do
next. Maybe it is in my coding for the logit model; I've not written that
into a program before and am not sure I am to write it for the
probability.
2)Even once that is solved, I still need to learn how to make the program
use frequency weights. Each line of data represents different numbers of
people. I collapsed a massive datafile this way. One line might refer to
7324 people, while another refers to 261. How do I make nlme use that
information to estimate the model appropriately?
3)How do I get nlme to read starting values from results of a logit model
estimated before the nlme run. i haven't tried that but assume those
starting values will be far better than my "all zeros" starting values,
which I am using just to debug the code. But,
4)I did not find a place I had to tell the program what the clustering was
on (states)? Is this unnecessary? And if it is necessary, how?
5)Well, that'll do for now I suppose. Any assistance is greatly
appreciated.
Sam
R : Copyright 2002, The R Development Core Team
Version 1.5.1 (2002-06-17)
R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type `license()' or `licence()' for distribution details.
R is a collaborative project with many contributors.
Type `contributors()' for more information.
Type `demo()' for some demos, `help()' for on-line help, or
`help.start()' for a HTML browser interface to help.
Type `q()' to quit R.
[Previously saved workspace restored]
> invisible(options(echo = TRUE))
> library(nlme)
Loading required package: nls
Loading required package: lattice
> fm1<-nlme(model=plfp ~ (exp(-1+bm + bmxd +
+ wm + wmxd +
+ bf + bfxd +
+ wf + wfxd +
+ yrsed + age + age2)) / (1+exp(-1+bm + bmrd +
+ wm + wmxd +
+ bf + bfxd +
+ wf + wfxd +
+ yrsed + age + age2)),
+ fixed=list(bm ~ 1, bmxd ~ 1,
+ wm ~ 1, wmxd ~ 1,
+ bf ~ 1, bfxd ~ 1,
+ wf ~ 1, wfxd ~ 1,
+ yrsed ~ 1, age ~ 1, age2 ~ 1,
+ data = PLFPANAL,
+ random=list(bm ~ 1, wm ~ 1, bf ~ 1, wf ~ 1),
+ start = c(bm = 0, bmxd = 0,
+ wm = 0, wmxd = 0,
+ bf = 0, bfxd = 0,
+ wf = 0, wfxd = 0,
+ yrsed = 0, age = 0, age2 = 0))
Error in parse(file, n, text, prompt) : parse error
Execution halted
* * * * * * * * * * * * * * * * * * * * * * * *
Samuel R. Lucas
Associate Professor of Sociology
University of California-Berkeley
410 Barrows Hall #1980
Berkeley, California 94720-1980
e-mail: Lucas at demog.berkeley.edu
home-page: http://sociology.berkeley.edu/faculty/lucas/
* * * * * * * * * * * * * * * * * * * * * * * *
More information about the R-help
mailing list