Table 1. Main results of the integrative logistic regression analyses.
| 2006 | 2009 | |||||
|---|---|---|---|---|---|---|
| Covariate | Parameter | Estimate | Posterior Probability | Estimate | Posterior Probability | |
| H’ (Host Shannon-Weiner Div) | α1 | -0.04 | 0.57 | -0.41 | 0.70 | |
| Mouse Relative Abundance | α2 | +0.52 | 0.96 | +0.66 | 1.00 | |
| NIPALL | Chipmunk Relative Abundance | α3 | +0.10 | 0.89 | — | — |
| Shrew Relative Abundance | α4 | +0.13 | 0.90 | -0.10 | 0.71 | |
| H’ (Host Shannon-Weiner Div) | γ1 | +0.20 | 0.74 | +1.67 | 0.71 | |
| Mouse Relative Abundance | γ2 | -0.73 | 0.96 | -1.93 | 0.90 | |
| NIPHIS | Chipmunk Relative Abundance | γ3 | -0.06 | 0.70 | -0.81 | 0.98 |
| Shrew Relative Abundance | γ4 | +0.26 | 0.95 | -0.57 | 0.89 | |
Estimates (i.e. posterior medians) of regression coefficients and posterior probabilities of positive or negative association between covariates and NIPAll or NIPHIS. A high posterior probability implies a high degree of confidence (little uncertainty) in the direction of the estimated association. For example, the second row under 2006 indicates a posterior probability of 0.96 (very high confidence) that mouse relative abundance is positively associated with NIPAll (slope estimate = +0.52). Missing entries correspond to a covariate that was omitted from our final 2009 model fit because there was negligible evidence of its association with NIPAll in the preliminary models that included all four covariates.