Disadvantages include possible lack of model fit, increased sample size requirements and the need for more specialised computer software [ 41].
The model fit increased modestly with the addition of diffuse OA (model C‐statistic 0.738, compared to 0.731 in the model excluding diffuse OA).
For example, item 1 had the second lowest factor loading in the PCA (though still substantial, Table 3), and model fit increased significantly in the CFA when the item was removed from the model.
Model fit increased with the 'agricultural land'/'natural vegetation' resistance ratio, but the strength of the relationship between genetic differentiation and IBR stabilized beyond a ratio 1000 1 (Table 2; Fig. 5).
The C-statistic, a measure of model fit, increased from 0.841 in model 1 (without any urinary proteins except for AER) to 0.857 with progranulin, 0.871 with THP, and 0.888 in the stepwise selection model, which included model 1 + THP, progranulin, and clusterin, and to 0.889 in the final model, which included also AGP.
The model exclusively including geographical distances between populations (i.e. IBD) was not significant and model fit increased with higher ratios of agricultural land/natural habitat resistance (Fig. 5), indicating that the subtle genetic differentiation observed within the study area probably reflects the impact of farming and land clearing on the species' demographic dynamics.
The model fit also increases substantially from the younger the older age group.
Inclusion of these risk factors marginally improves the model fit but increases the AIC value, so the models are rejected as not parsimonious.
We expect that the threshold for a significant model fit should increase monotonically with the model complexity.
These indicators show a substantial improvement of the model fit with increasing number of groups of cDNAs.
