**Supplemental Tables 1-4: Model Comparisons for Individual Data Sets **

These tables (Microsoft Excel format) provide various model comparisons using the AIC, BIC, and GLRT. Comparisons are presented for various nested variant/model pairs. For each data set and the overall pooled data set, the tables provide the ln likelihoods of the best model fits, the number of data points (needed to compute the BIC), and the number of model parameters. The tables then do pairwise model comparisons based on AIC scores, BIC scores, and the GLRT (at the p-value specified in the table.) Each table is set up as a spreadsheet, so that as many quantites as possible are computed from other quantities in the table. Entered data is highlighted in light blue, computed comparison are highlighted in light purple. The interpretation of "preferred model" for AIC and BIC is simply that the preferred model had the better score. For the GLRT, the complex model is preferred if the simple model is rejected at the specified p-value and the simple model is preferred otherwise.

Supplemental Table 1. Comparisons of the CM, IA, and IC variants of the {g,M} model. The table provides ln likelihoods, parameter counts, and pairwise model comparisons based on the BIC, AIC, and GLRT.

Supplemental Table 2. Comparisons of the CM, IA, and IC variants of the {g,q,M} model. The table provides ln likelihoods, parameter counts, and pairwise model comparisons based on the BIC, AIC, and GLRT.

Supplemental Table 3. Comparisons of CM variant of {g,M} and {g,q,M} models. The table provides ln likelihoods, parameter counts, and pairwise model comparisons based on the BIC, AIC, and GLRT.

Supplemental Table 4. Comparisons of IA variant of {g,M} and {g,q,M} models. The table provides ln likelihoods, parameter counts, and pairwise model comparisons based on the BIC, AIC, and GLRT.