Goodness of fit of the dual trajectory models with differing numbers of clusters
| Clusters | AIC | BIC | Adjusted BIC | Smallest class (overall) (%) | AvePP* (range) |
| 1 | 23 241.43 | 23 277.38 | 23 251.98 | ||
| 2 | 17 999.11 | 18 066.52 | 18 018.89 | ||
| 3 | 16 185.44 | 16 284.30 | 16 214.45 | 16 | 0.98–0.98 |
| 4 | 15 088.50 | 15 218.82 | 15 126.74 | 15 | 0.95–0.99 |
| 5 | 14 310.02 | 14 471.80 | 14 357.50 | 9 | 0.94–0.99 |
| 6 | 14 018.94 | 14 212.17 | 14 075.64 | 7 | 0.93–0.98 |
| 7 | 13 820.48 | 14 045.17 | 13 886.42 | 4 | 0.92–0.98 |
| 8 | 13 701.00 | 13 957.15 | 13 776.17 | 3 | 0.89–0.99 |
| 9 | 13 628.19 | 13 915.79 | 13 712.59 | 0.1 | 0.88–1.00 |
*AvePP for participants allocated to cluster.
AIC, Akaike information criterion; AvePP, average posterior probability; BIC, Bayesian information criterion.