[R-sig-ME] lme varFix under ML fit does not match coefficients standard error
Ben Bolker
bbo|ker @end|ng |rom gm@||@com
Sun Feb 18 20:52:25 CET 2024
From ?summary.lme:
adjustSigma: an optional logical value. If ‘TRUE’ and the estimation
method used to obtain ‘object’ was maximum likelihood, the
residual standard error is multiplied by sqrt(nobs/(nobs -
npar)), converting it to a REML-like estimate. This argument
is only used when a single fitted object is passed to the
function. Default is ‘TRUE’.
On 2024-02-18 2:34 p.m., Vaida, Florin via R-sig-mixed-models wrote:
> Hello all,
>
> This is probably known, but it’s news to me: the standard errors printed for the lme model fit run under method=”ML” are in fact those computed under method=”REML”.
> Is this the expected behavior? And if so, are there any reasons for this choice?
> Reproducible example below.
>
> Thanks,
> Florin
>
> library(nlme)
> fit.reml = lme(log(conc) ~ Time, random=~1|Subject, data=Glucose, na.action=na.omit, method="REML")
> (se.reml = summary(fit.reml)$tTable[,2])
> ## (Intercept) Time
> ## 0.019457141 0.005829144
> (se.reml = sqrt(diag(summary(fit.reml)$varFix)))
> ## (Intercept) Time
> ## 0.019457141 0.005829144
> fit.ml = lme(log(conc) ~ Time, random=~1|Subject, data=Glucose, na.action=na.omit, method="ML")
> (se.ml = summary(fit.ml)$tTable[,2]) # they match the REML SE’s
> ## (Intercept) Time
> ## 0.019457141 0.005829144
> (se.ml = sqrt(diag(summary(fit.ml)$varFix))) # they do not match the tTable SE’s
> ## (Intercept) Time
> ## 0.019405324 0.005813621
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
More information about the R-sig-mixed-models
mailing list