We performed stepwise modeling, retaining variables that were significant at P<0.05 or that were deemed a priori to be clinically important in the multivariable model.
The Health Belief Model was used as the theoretical framework for variable selection and modeling, particularly looking at variables that fit into the model's four main constructs of perceived susceptibility, severity, barriers and benefits [ 23].
In further modeling exercises, we considered variables that could alter the observed association between segregation and land cover characteristics through confounding or mediation.
In exact model reduction [ 4, 5] and layer-based modeling [ 3] there exist variables that correspond to sums of species of a corresponding conventional model.
Multivariate multilevel modeling (adjusted) (Table 2) showed that, after controlling for the variables that were significant in bivariate multilevel modeling.
Thus, it is necessary to limit the number of free variables that are inferred, leading us to consider a modeling approach that builds on the existing domain knowledge.
HVD modeling estimates log-normal distributions from data on pharmacokinetic and pharmacodynamic variables that affect individual sensitivities to the toxic response.
Advantages of the approach include reduction of potential confounding by variables that are typically associated with mortality (e.g., smoking history, blood pressure), simplification of modeling, and reduction of vulnerability to model mis-specification bias (Armstrong 2003; Schwartz 2005b).
From the modeling perspective, VSTLF models can rely on lagged load as an independent variable in addition to others such as weather and calendar variables that are commonly used in STLF.
Structural equation models provide a latent variable modeling framework that is useful in detailed examinations of dimensionality.
