Error with nlme

Question

For IGF data from nlme library, I\'m getting this error message:

lme(conc ~ 1, data=IGF, random=~age|Lot)
Error in lme.formula(conc ~ 1         


        

Answer 1

If you plot the data, you can see that there is no effect of age, so it seems strange to be trying to fit a random effect of age in spite of this. No wonder it is not converging.

library(nlme)
library(ggplot2)

dev.new(width=6, height=3)
qplot(age, conc, data=IGF) + facet_wrap(~Lot, nrow=2) + geom_smooth(method='lm')

I think what you want to do is model a random effect of Lot on the intercept. We can try including age as a fixed effect, but we'll see that it is not significant and can be thrown out:

> summary(lme(conc ~ 1 + age, data=IGF, random=~1|Lot))
Linear mixed-effects model fit by REML
 Data: IGF 
       AIC      BIC    logLik
  604.8711 618.7094 -298.4355

Random effects:
 Formula: ~1 | Lot
        (Intercept) Residual
StdDev:  0.07153912 0.829998

Fixed effects: conc ~ 1 + age 
                Value  Std.Error  DF  t-value p-value
(Intercept)  5.354435 0.10619982 226 50.41849  0.0000
age         -0.000817 0.00396984 226 -0.20587  0.8371
 Correlation: 
    (Intr)
age -0.828

Standardized Within-Group Residuals:
        Min          Q1         Med          Q3         Max 
-5.46774548 -0.43073893 -0.01519143  0.30336310  5.28952876 

Number of Observations: 237
Number of Groups: 10