Hello, dear BUGS-users!
I am compiling a metaanalysis over a set of studies which can be partitioned into three classes according to the setting in which they were done. Target quantity are treatment effects which are measured by a difference in pre-postintervention proportions. My model is a Bayesian Hierachical Normal Model with binominal approximation of the study-(lower)-level variances and uninformative gamma hyperpriors. Individual models were computed for the complete set of studies and the three subsets.
Now the following strange results pop up: The population-level (prior) expectation for the COMPLETE set of studies is higher than for any of the three mentioned subsets! (see below) Is this an effect of the assumption of study-data integration via the normal prior? Intuitively, I would interpret the expectation of the all-studies-model as a kind of "mean" effect with regard to the subset models. Where is the mistake?
Any help would be greatly appreciated!
Kind regards:
Gero
----
Studies Treatment effect (d’) SD(d’) VAR(d’) SD(Var(d’))
(mu.theta) (sigma.theta)
All 0,068113 0,008764 0,09146 0,00069
Type 1 0,045636 0,014612 0,08217 0,01277
Type 2 0,053449 0,027717 0,12511 0,02366
Type 3 0,045630 0,009861 0,07108 0,00896
----
_______________________________________________________________
SMS schreiben mit WEB.DE FreeMail - einfach, schnell und
kostenguenstig. Jetzt gleich testen! http://f.web.de/?mc=021192
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|