Table 1.
Structural equation model and model descriptors by model (with both latitude and temperature, only latitude or temperature, or neither) ranked by model delta AIC score.
| SEM Model | χ 2 | df | p-val | Comparative Fit Index | Root Mean Square Error of Approximation | Standardized Root Mean Square Residual | Adjusted Goodness of fit | AIC | ΔAIC |
|---|---|---|---|---|---|---|---|---|---|
| Temperature | 1.804 | 3 | 0.614 | 1 | 0.000 | 0.016 | 0.894 | 906 | 0 |
| Latitude | 1.882 | 3 | 0.597 | 1 | 0.000 | 0.016 | 0.884 | 914 | 7 |
| None | 2.328 | 3 | 0.507 | 1 | 0.000 | 0.019 | 0.877 | 932 | 26 |
| Temp + Lat | 1.896 | 4 | 0.755 | 1 | 0.000 | 0.018 | 0.989 | 963 | 57 |
I report model chi-square values (χ2), degrees of freedom (df), p values (here, the larger the better), comparative fit square values (the closer to 1 the better), root mean square errors of approximation (the closer to 0, the better, values above 0.08 are suggestive of a bad fit), standardized root mean square residuals (smaller than 0.08 suggest a good fit), adjusted goodness of fit (can be interpreted as the proportion of explained variance), Aikaike Information Citerion values (AIC) and delta AICs (ΔAIC). For full model descriptions see Table S2 in Appendix.