Table 2. Results for metrics from an information criteria based analysis for four different models.
The lower value means a better fit. Log likelihood (logLik) and deviance are metrics used for computing the AIC.
| logLik | deviance | AIC | BIC | df.residual | ||
|---|---|---|---|---|---|---|
|
Modeling 2:1 ratios |
Full model | -6.11 | 0.02 | 20.22 | 48.67 | 9059 |
| Model without Size Difference | -36.16 | 44.68 | 78.31 | 99.65 | 9060 | |
| Model without Height Ratio | -285.84 | 541.85 | 577.68 | 599.02 | 9060 | |
| Marker only Model | -285.84 | 541.84 | 575.69 | 589.91 | 9061 | |
|
Modeling 1:2 ratios |
Full model | -7.13 | 0.02 | 22.26 | 50.70 | 9052 |
| Model without Size Difference | -10.72 | 4.00 | 27.44 | 48.77 | 9053 | |
| Model without Height Ratio | -260.16 | 504.70 | 526.31 | 547.64 | 9053 | |
| Marker only model | -260.16 | 504.67 | 524.31 | 538.53 | 9054 | |