The recorded concentrations (of IgG1, IgG2, IgM, IgA and BSA) were analysed according to a statistical model including the fixed effects of sampling occasion, fluid and the interaction between sampling occasion and fluid.
These ratios were analysed according to a statistical model including the fixed effects of sampling occasion, parameter (IgG1, IgG2, IgM, IgA and BSA) and the interaction between sampling occasion and parameter.
For the measurements of Salmonella in the flock 1 week before the end of rearing, the time percentages were approximate as they included the week between this sampling occasion and bird harvest.
For pairwise analyses, the correlation coefficient for each pair of sampling occasions provides an estimate of repeatability over the time interval between sampling occasions (Lessells and Boag 1987).
The univariate between-subject effects of significant MANOVA factors revealed that chlorophyll-a concentrations were affected by the resource treatments (F1,27 = 8.906, p = 0.006), fish density (F4,27 = 3.278, p = 0.026) and differed between sampling occasions (F3,27 = 8.794, p<0.001; Figure 1).
They revealed that total cladoceran densities depended on resource level (F1,27 = 14.659, p = 0.001) and fish density (F4,27 = 15.752, p<0.001), and that cladoceran densities differed between sampling occasions (F3,27 = 7.525, p = 0.001).
Moreover this approach evaluates the relative direction and size of effects of factors on dependent variables within, rather than across, block units, allowing for evaluation of factor effects even if dependent variable levels differ between sampling occasions [38].
Moreover, all treatment groups showed the same pattern in time, ie. no difference between sampling occasions (p = 0.572).
However, the time intervals between the samplings in the current studies (1, 2 and 6 months) were sufficient for the dogs to be re-infected between sampling occasions.
The numbers of lymphocytes did not differ significantly between sampling occasions, but was lower than the normal range (2.5 7.5 × 10/l) in 8 cows at calving.
Further, the different lengths of period at risk for the incidence estimates at the second and third sampling occasions was not optimal, because it made comparisons between sampling occasions impossible.
