Table 1. Model comparison based on the log-likelihood and Akaike Information Criterion (AIC; smaller is better).
Model | Predictor(s) | Cov | Log-likelihood | AIC |
---|---|---|---|---|
1 | logit(PANC) | Ind | −23.96 | 53.92 |
2 | logit(PANC) + Formal | Ind | −23.42 | 54.85 |
3 | logit(PANC) + DR | Ind | −22.37 | 52.73 |
4 | logit(PANC) + SEQ | Ind | −22.84 | 59.68 |
5 | logit(PANC) + MMR | Ind | −23.75 | 55.51 |
6 | logit(PANC) + Formal + DR | Ind | −22.36 | 54.73 |
7 | logit(PANC) + Formal + DR + MMR | Ind | −22.34 | 56.67 |
8 | logit(PANC) + Formal + DR + MMR + HFB | Ind | −22.20 | 58.39 |
9 | Formal + DR + MMR + HFB | Ind | −37.81 | 87.63 |
10 | logit(PANC) | SAR | −23.95 | 55.91 |
11 | logit(PANC) + DR | SAR | 22.37 | 54.73 |
12 | logit(PANC) + Formal + DR + MMR + HFB | SAR | −31.80 | 77.61 |