A full model, with proportions specific to each site-time combination (S.T), gave an even lower AIC value, but the AIC difference between this and the S+T model was modest, indicating that the S+T model (Model D) provides an adequate description of the data on apparent prevalence.
These results suggest that, globally, the differences produced by each model are modest.
Yet, as we have argued elsewhere, these responses, which might be considered the 'standard' model, are modest given the scale of the challenge [ 9].
The discriminatory power of the model was modest as determined by area under the receiver operating characteristic (ROC) curve (c = 0.65).
The goodness-of-fit of this model was modest (Cox and Snell R2 = 0.21, Nagelkerke R2 = 0.42, chi-square = 914.2, P < 0.0001).
Thus, the magnitude of inflammatory changes in our model was modest, similar to that we previously observed in the early-life model of asthma following pneumovirus infection (Siegle et al., 2010).
The degree of inhibition in tumor formation by MSC in our mouse model was modest compared with that demonstrated in chemical carcinogen-induced rat mammary tumor models [ 3, 4].
DOI: http://dx.doi.org/10.7554/eLife.05241.006 The improvement in predictive power of the subunit model over the LN model was modest, in part because the predictions were performed over a large ensemble of random noise stimuli that only infrequently and weakly activate cones and subunits in a manner that differentiates the model predictions.
Thus, since the difference between the non-post-stratified model and the post-strata models were modest, we would rather recommend the non-post-stratified model (which disregards the land use and vegetation types) to be more adequate for most applications that will involve large-scale AGB estimation supported by ALS data, at least until high quality thematic maps are made available.
Whereas the improvement in the models was either statistically significant (comparing health center and nonspatial models) or trending toward significance (comparing spatial and nonspatial models), the absolute differences in the AUCs from spatial and nonspatial models were modest.
