The GLMM procedure enables individual identity to be accounted for (by entering it as a random factor), although this does not completely address the problem of potential inter-dependency in the individuals' foraging decisions.
Here, we accounted for the nested nature of our data by including species as a random factor.
We adjusted the analysis by care home (as a random factor), baseline Barthel index score, and stratification factors: trial administrative centre (11 centres) and type of care home (residential or nursing).
GLMMs can be viewed as an extension of general linear models accounting for repeated measurement of the same individual by including subject as a random factor in the model (Pinheiro and Bates 2000).
To investigate the impact of errors in the observed metabolite concentrations on the predicted flux distribution the concentration values given in figure 2 were perturbed by multiplying them with a random factor obeying an exponential normal distribution with controlled standard deviation (see caption of figure 3).
The interobserver variability in quantitative MCE was assessed by computing intraclass correlation coefficients using one-way ANOVA with a random factor and by calculating the limits of agreement between measurements made by use of Bland-Altman analysis [ 17].
Within-make, plate-to-plate variation was accounted for by including plate ID as a random factor (Equation 2).
Like the previous model, the analysis includes Trial nested within Block as a random factor and two by-subject random slopes, for Trial and for Block.
Additionally, the analysis includes Trial nested within Block as a random factor and two by-subject random slopes, for Trial and for Block.
Since eggs were collected from each female five times during the sample period, we included female identity as a random factor (shrunk by REML estimation) to account for non-independence of within-female measures.
Statistical analysis was performed using mixed-effects ANOVA, with dosage as a fixed factor and donor identity as a random factor, followed by post hoc Dunnett t-tests.
