Table 1. Latent class growth analysis procedure and model fit.

Number of latent subgroups	2	3	4	5	6	7	8	9
AIC*	49727	49137	48957	48838	48787	48745	48722	48712
BIC*	49774	49202	49040	48938	48905	48880	48875	48883
Adjusted BIC*	49749	49167	48995	48885	48841	48807	48793	48790
Entropy	74%	74%	71%	75%	76%	75%	68%	69%
Lo, Mendell, Rubin LRT*	2 vs. 1 1782 P<0.00001	3 vs. 2 572 P<0.00001	4 vs. 3 178 P = 0.0010	5 vs. 4 120 P<0.00001	6 vs. 5 55 P = 0.0118	7 vs. 6 46 P = 0.0169	8 vs. 7 28 P = 0.3039	9 vs. 8 33 P = 0.3255
Bootstrapped LRT*	P<0.00001	P<0.00001	P<0.00001	P<0.00001	P<0.00001	P<0.00001	P<0.00001	P<0.00001
N for each class	1:N = 1792 (67%) 2:N = 868 (33%)	1:N = 1105(42%) 2:N = 1292 (48%) 3:N = 263 (10%)	1:N = 1234 (46%) 2:N = 110 (4%) 3:N = 562 (21%) 4:N = 754 (28%)	1:N = 1247 (47%) 2:N = 52 (2%) 3:N = 108 (4%) 4:N = 739 (28%) 5:N = 514 (19%)	1: N = 709 (27%) 2: N = 1208 (45%) 3: N = 116 (4%) 4: N = 42 (2%) 5: N = 528 (20%) 6: N = 57 (2%)	1: N = 667 (25%) 2: N = 526 (20%) 3: N = 39 (1%) 4: N = 47 (2%) 5: N = 145 (5%) 6: N = 73 (3%) 7: N = 1163 (44%)	1: N = 570 (21%) 2: N = 426 (16%) 3: N = 76 (3%) 4: N = 688 (26%) 5: N = 38 (1%) 6: N = 145 (5%) 7: N = 49 (2%) 8: N = 668 (25%)	1: N = 553 (21%) 2: N = 583 (22%) 3: N = 40 (25%) 4: N = 661 (25%) 5: N = 47 (2%) 6: N = 14 (1%) 7: N = 89 (3%) 8: N = 514 (19%) 9: N = 159 (6%)

* Performance measures: AIC: Akaike information criterion; BIC: Bayes information criterion; LRT: Likelihood Ratio Test