The first stage involves identifying the possible factors and their interactions, constructing a partial factorial design to allow the impact of first- and second-order interactions on the performance measures to be identified, and analysing the resultant performance measures using generalised linear models.
As blood samples were taken at different times postoperatively as repeated measures, a generalised linear mixed effect model (GLMM) was used to investigate the extent to which blood Co levels changed in relation to the time since the original hip replacement operation, with patient age and sex, the device used and bearing diameter of the implant as explanatory variables.
Repeated measures analyses using generalised linear mixed models which take into account the covariance pattern of the data will be used to investigate changes over time in the major outcomes and whether these changes differ between the intervention and control groups.
Differences in baseline-reported symptoms between current and former hairdressers were measured using the generalised linear model for binomial data (glm command in Stata with identity link) (giving PD and 95% CI).
For each colony and immune measure we fit generalised linear models with principal components that explained ≥ 5%% of the variation as response variables to homozygosity weighted by locus (HL) [ 52], sex and their interaction as explanatory terms, removing interactions if they were non-significant [ 71].
Instead, since the posterior contains information on which parameters are more likely than others, a measure of central tendency (preferably the mode) could be used and passed into standard frequentist generalised linear models, particularly if the posterior distributions of interest are precise and not highly skewed.
Performance measures were compared using generalised linear mixed models, accounting for inter-reader performance heterogeneity (P < 0.05).
Because each individual chicken was included in the data set many times, we corrected for repeated measures by fitting a generalised linear mixed model (GLMM procedure), nesting day within individual chicken in the random model.
For person-based analysis, relative risks (RRs) with 95% confidence intervals (95% CI) of SCC incidence for each categorical physical activity measure were estimated by generalised linear models, specifying Poisson distribution with a robust error variance [ 28] and person-years of follow-up as offset.
Longitudinal (or serially measured) endpoints will be evaluated by generalised linear models and linear mixed.
