TABLE 2.
Generalized linear model with gamma residuals (link = log) predicting recovery of 11 viruses in the liquid phasea
| Variable | Model estimate |
|---|---|
| Intercept | 3.759 (0.078)*** |
| Vol | –0.007 (0.001)*** |
| Method: AS | –1.245 (0.092)*** |
| Method: BE-PEG | –0.368 (0.092)*** |
| Method: IP | –1.55 (0.101)*** |
| Method: PEG | –1.569 (0.094)*** |
| Water type: WW | –0.848 (0.057)*** |
| AICb | 4,748.24 |
| R squared | 0.597 |
| Adjusted R squared | 0.594 |
a
Pellet methods were excluded from the model due to nonstandard starting volumes. The model results include the variable coefficient on a log scale indicating its effect (positive numbers indicating increased recovery), followed by the standard error in parentheses and a significance code (***, P < 0.001) rounded to three decimal places. The Amicon method and deionized water (DW) are predicted using the intercept.
b
AIC, Akaike information criterion.