Table 1.
Model fit statistics for estimated latent class models for COVID-19 protective and risky behaviors.
| N Classes | Log-likelihood | Approx. BFk,k+1 | BIC | cAIC | AWE |
|---|---|---|---|---|---|
| 1 | −18030.95 | <0.10 | 36546.02 | 36618.02 | 37246.13 |
| 2 | −17194.38 | <0.10 | 35363.72 | 35508.72 | 36773.68 |
| 3 | −16695.06 | <0.10 | 34855.91 | 35073.91 | 36975.70 |
| 4 | −16435.16 | >10 | 34826.96 | 35117.96 | 37656.59 |
| 5 | −16200.59 | >10 | 34848.65 | 35212.65 | 38388.13 |
| 6 | −16009.01 | >10 | 34956.33 | 35393.33 | 39205.65 |
| 7 | −15890.35 | >10 | 35209.85 | 35719.85 | 40169.00 |
| 8 | −15749.19 | – | 35418.37 | 36001.37 | 41087.36 |
Approx. BF = Approximate Bayes Factor, BIC = Bayesian information criterion, cAIC = consistent Akaike’s Information Criterion, AWE = Approximate weight of evidence criterion