[R-sig-ME] mixed effects model glmer
westm490 at gmail.com
Wed Sep 23 20:36:05 CEST 2015
I am trying to fit a mixed effects model with repeated measures data.
y variable = percentage (# females/total)
x variable = percentage
measured across multiple sites for 4 years.
here's the model:
y <- cbind(total females, (Total - total females)))
mod1 <- with(data, glmer(y ~ disease prevalence + (1|Site) + (1|Year),
family = binomial, data = data1))
1) This model runs, but the summary(mod1) just generates a series of the
following....which doesn't make any sense so something must be wrong with
the model specification...I'm just not sure what.
2) Also, what is the default AR correlation on these models (i.e., do I
need to specify it or is the temporal psuedoreplication taken care of)?
3) Finally, do you suggest another form of the model that's better etc.?
Estimate Std. Error z value
(Intercept) -1.60267 0.11618 -13.794 <
disease prevalence -0.40212 0.15557 -2.585 0.009745 **
disease prevalence 0.035088231 -1.46452 0.22860 -6.406 1.49e-10
disease prevalence 0.064935065 -0.36344 0.30810 -1.180 0.238157
disease prevalence 0.078507945 -2.57479 0.46537 -5.533 3.15e-08
disease prevalence 0.120039255 -3.30998 0.71915 -4.603 4.17e-06
disease prevalence 0.182623706 -0.14362 0.19899 -0.722 0.470438
Many thanks in advance,
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models