Table 2.

Model fit statistics for linear zero-Inflated Poisson and negative binomial growth mixture models of mania symptoms

N classes	Model type	AIC	BIC	aBIC	Entropy	Smallest class size	Average posterior probabilities	VLMR-LRT p-value
1	Poisson	105,679.68	105,708.71	105,695.99	1.00	100%	1.00	NA
	Negative Binomial	89,987.86	90,038.66	90,016.41	1.00	100%	1.00	NA
2	Poisson	91,833.39	91,884.18	91,861.94	0.699	13%	0.93–0.93	0.0125
	Negative Binomial	84,164.59	84,237.16	84,205.38	0.750	36%	0.92–0.95	0.0000
3	Poisson	85,915.88	85,988.45	85,956.67	0.821	9%	0.90–0.94	0.0004
	Negative Binomial	82,765.81	82,860.15	82,818.84	0.710	11%	0.85–0.90	0.0654
4	Poisson	84,657.54	84,751.87	84,710.56	0.776	4.2%	0.83–0.92	0.0688
	Negative Binomial	82,597.49	82,713.59	82,662.75	0.630	6.6%	0.67–0.90	0.3115
5	Poisson	84,079.03	84,195.13	84,144.29	0.782	3.6%	0.77–0.92	0.2152
	Negative Binomial	82,529.11	82,666.99	82,606.61	0.666	1.1%	0.63–0.85	0.4940
6	Poisson	83,826.34	83,964.21	83,903.83	0.768	1.7%	0.77–0.90	0.4015
	Negative Binomial	82,508.43	82,668.08	82,598.16	0.585	1.5%	0.58–0.83	0.4641

Note. AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; aBIC, sample-size adjusted Bayesian Information Criterion; VLMR-LRT, Vuong-Lo-Mendell-Rubin likelihood ratio test. Bolded values indicate the better-fitting model (Poisson or negative binomial) for comparisons with the same number of classes