To obtain p-values from mixed models, we conducted a likelihood ratio (LR) test between each full model (with treatment and random intercepts effects) and a nested, null case of the model (random intercept effects only).
Finally, based on realistic US maps and traffic models, we conducted simulations to evaluate the feasibility and efficiency of our methods.
Using climate projections from an ensemble of 16 models, we conducted an assessment of a midcontinental area of North America Minnesotaa) based on a resistance, resilience, and facilitation framework.
To test these models, we conducted extensive studies in 1992, and again in 2002.
In order to overcome general constraints of the quality of linear regression models, we conducted tests for heteroscedasticity, serial correlation, multi-collinearity, endogeneity and normal distribution.
Using both slip models, we conducted a nonlinear tsunami simulation using high-resolution bathymetry data to analyze the tsunami propagation around Sendai Coast.
For each of the final multivariable and interaction models, we conducted 2 sensitivity analyses.
To test the models, we conducted an eye-tracking experiment using a wide variety of different images.
To assess aim 3 (the value of using the TREC care unit definition in planned explanatory models), we conducted multi-level modeling (Hierarchical Linear Modeling, HLM).
In order to see if we can improve the performance further by adding more features to the models, we conducted an additional test as shown in Table 5.
