[R-sig-ME] Variance of the random effects

[Hester Lingsma]

Dear R users,
In a model with a random intercept for center, I use the variance of the 
random effect to express the differenrences between centers in a 95% 
range of odds ratios. I use:
exp(1.96*(sqrt(as.numeric(VarCorr(random1)[[1]]))))
exp(1.96*(-(sqrt(as.numeric(VarCorr(random1)[[1]])))))
If I look at the posterior estimates of the individual centers, I find 
out that (much) more than 5% is actually out of the range I calculated. 
Am I making a mistake?
Thanks,
Hester

-- 
_________________________________________________
Hester F. Lingsma, MSc
Dept of Public Health
Room AE-141
Erasmus MC
P.O. Box 2040
3000 CA Rotterdam
The Netherlands
Phone: (+31) (0)10 7038458/7038460
Mobile: (+31) (0)6 26467338