Table 2.
Model selection of the growth experiment based on Akaike information criteria (AIC). The best fit model estimating bacterial count of the strain (B067 or B398) within time is underlined. p value indicates the significance of the term if removed from the higher model
| Model | AIC | df | p |
|---|---|---|---|
| S+NL+A+S:NL+S:A+NL:A+S:NL:A+(1 | ID) | 3,257.8 | 26 | |
| S+NL+A+S:NL+S:A+NL:A+(1 | ID) | 3,289.0 | 21 | <.001 |
| S+NL+A+S:A+NL:A+(1 | ID) | 3,294.4 | 20 | .007 |
| S+NL+A+S:A+(1 | ID) | 3,364.5 | 11 | <.001 |
| S+NL+A+ (1 | ID) | 3,365.6 | 10 | <.001 |
S, strain; NL, nutrient level; A, day; and ID, culture replicate (random variable). + describes the main effects, and colon describes the interactions.