[S] Mixed Model with AR(1) errors, variances

Regula Schwager-Suter (suter@inw.agrl.ethz.ch)
Mon, 28 Sep 1998 15:19:53 +0200


I am using S-Plus, Version 3.4 Release 1 for Sun SPARC, SunOS 5.3

I have repeated measures data on several animals (not all of the animals
have the same number of measurements). Because the data are collected over
time, there could be correlation between serial observations on the same
experimental unit.
So I used the following model in S-Plus to analyze my data:

> model_lme(fixed = Y ~ C(effect1, treatment) + C(effect2, treatment) +
+ C(effect3, treatment) + covariate1 + covariate2 + covariate3 + ... +
+ covariate8, random= ~1, cluster= ~subject, serial.structure="ar1",
+ serial.covariate= ~week)

The within-individual errors are assumed to follow the AR(1)model as
described in Chi and Reinsel (1989):
u(t) = phi*u(t-1) + e(t) e(t)~N(0, sigma^2) t = time of data collection

The S-Plus output delivers the variance of the u(t). I would like to know
the variance of the e(t)'s, too. Can I do this and if yes, how can I do this?

Thank you very much for your help.

Best regards
Regula
_____________________________________

Regula Schwager-Suter
Swiss Federal Institute of Technology
Institute of Animal Science
Animal Breeding
Clausiusstrasse 50
CH-8092 Zürich
phone: +41 (0)1 632 39 56
fax: +41 (0)1 632 12 60
e-mail: suter@inw.agrl.ethz.ch
http://www.tz.inw.agrl.ethz.ch/~suter
_____________________________________
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news