Table 1.

Model fit statistics for quadratic LCGMM comparing solutions for 1 to 7 latent classes

Sex	Latent classes	BICa	LMR-LRTb (P-value)	BLRTc (P-value)	Convergence	Entropy value	Log Likelihood	Number of parameters
Boys	1	23605.4	N/A	N/A	Yes	N/A	−11738.3	19
	2	22846.6	0.03	0.00	Yes	0.98	−11345.4	23
	3 d	22644.4	0.13	0.67	Yes	0.99	−11230.7	27
	4	22544.7	0.53	1.00	Yes	0.98	−11167.3	31
	5	22476.5	0.77	0.67	Yes	0.93	−11119.7	35
	6	22438.4	0.61	1.00	Yes	0.94	−11087.1	39
	7				NO			43
Girls	1	27949.8	N/A	N/A	Yes	N/A	−13909.8	19
	2	27226.2	0.00	0.00	Yes	0.94	−13534.3	23
	3	27155.0	0.10	0.00	Yes	0.76	−13485.0	27
	4 d	27083.9	0.23	0.00	Yes	0.80	−13435.7	31
	5	27026.0	0.45	0.6	Yes	0.82	−13393.1	35
	6	26990.9	0.63	1.00	Yes	0.78	−13361.8	39
	7	26961.2	0.27	1.00	Yes	0.80	−13333.3	43

LCGMM latent class growth mixture modeling

^aBayesian Information Criterion

^bLo, Mendell and Rubin Likelihood Ratio Test

^cBootstrapped Likelihood Ratio Test, N/A – not applicable

^dThe optimal class solutions according to the model fit criteria are shown in bold