James,
I follow this question with interest because i have used these menus quite a lot, and other people here far more knowledgeable than me will answer, however:
1) if the second, third timepoint are somewhat dependent on the starting point i generally use antedependence OR if a full, balanced dataset for all factors then i go with default option and they all work well
2) i have generally also used AIC, and someone once said that the most minimal value will describe the best model, but a net change of <50 meant very little effect of altering the fixed/random effects. Of course, the size of change is related to the original number, but <10 wouldnt really be of much concern about which model was best fit.
i have rarely specified additional random variation, and so i also would be interested in the answer to that...i daren not answer before the professionals...
david.
########################################################################
To unsubscribe from the GENSTAT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=GENSTAT&A=1
This message was issued to members of www.jiscmail.ac.uk/GENSTAT, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
|