Table 2.

Fit statistics for competing variance components models for the impulsive–compulsive, impulsivity, and obsessiveness phenotypes.

Phenotype	Model	− 2*LL	df	χ2	P	BIC	Hi (Pr | D)
Impulsive–compulsivea	ACE	− 437.897	12	28.101	0.141	906.714	0.055
	AE	− 439.103	13	19.639	0.105	903.973	0.217
	CEd	− 437.897	13	17.226	0.189	901.560	0.727
	E	− 447.563	14	36.558	0.001	915.739	0.001
Impulsivity	ACE	− 426.603	5	5.177	0.395	878.972	0.046
	AEd	− 426.603	6	5.177	0.521	873.819	0.604
	ADE	− 424.014	5	3.699	0.594	877.494	0.096
	CE	− 428.187	6	8.346	0.214	876.987	0.124
	E	− 430.712	7	13.396	0.063	876.884	0.130
Obsessiveness	ACE	− 407.493	5	7.404	0.192	840.752	0.030
	AEb	− 407.582	6	7.583	0.270	835.777	0.356
	CEc	− 407.504	6	7.426	0.283	835.621	0.385
	E	− 410.602	7	13.622	0.058	836.663	0.229

− 2*LL − 2 × log likelihood, df degrees of freedom for the chi square test statistic, χ² chi square test of model fit, p probability of the chi square test statistic, BIC Bayesian information criterion, H_i(Pr|D) Bayesian conditional posterior probability of model H_i compared with H_k models, A additive genetic effects, C common environment effects, E unique environment effects + error variance, D dominance genetic effects.

^aAge included in all models as a regressor for the impulsive-compulsive  phenotype.

^bBayesian posterior probability of AE compared to CE = [H₀ (Pr | D) = 0.481].

^cBayesian posterior probability of CE compared to AE = [H₁ (Pr | D) = 0.519].

^dPreferred model based on weight of evidence.