st: Guidance on predict in longitudinal data analyses

[Anthony Staines <anthony.staines@ucd.ie>]

Hi,
I have two questions - one is a Stata question, the second is a more 
general statistical question.

I am working on a clinical trial of exercise in pregnancy, where a mjor 
interest is in describing the temporal changes in physiological variable 
measured during and after pregnancy.

I want to summarise the results of some models by presenting the 
modelled values of physiological variables at 6 weeks and 36 weks of 
gestation like this :-
pontoon
Variable 
	6 weeks 
		36 weeks
PEFR 
		500			450	
V0 
		2.3			1.9
GET 
		0.75			0.4
 etc.. (The numbers are invented).

Stata question
How can I use 'predict' to generate values (and confidence intervals) 
for model predictions at specifc values of the time variable?
I don't want to generate a datset with the mean values of the predictor 
variables at every timepoint, unless I have to.

More general question.
For these particular data random variable models didn't fit any better 
than simpler models - but it made me think, and I have no idea what a 
'prediction' for a random variables model means - does it include some 
estimate of the 'mean' values of the random variable? Any guidance?

Best wishes,
Anthony Staines
--
Anthony Staines, Lecturer in Epidemiology, Department of Public Health,
  UCD, Earlsfort Terrace, Dublin2, Ireland. http://phm.ucd.ie/
  E-mail :- Anthony.Staines@ucd.ie Web :- http://astaines.ucd.ie/
    Work :- + 353 1 716 7345       Fax :- +353 1 716 7407
    Home :- + 353 1 849 1089    Mobile :- +353 1 86 606 9713	


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/