Table 3.

Odds ratios and 95 % confidence intervals based on final generalized estimating equation models of predictors of substance dependence (n = 743)

Predictor	Cannabis	Sedative	Stimulant	Cocaine	Opioid	Hallucinogen	Any
Gender	3.64*	0.84	4.60*	2.72	1.13	3.73	1.15
	(1.33–9.88)	(0.30–2.39)	(1.11–19.08)	(0.80–9.19)	(0.57–2.25)	(0.93–14.90)	(0.56–2.38)
	p = .011	p = .75	p = .04	p = .11	p = .73	p = .06	p = .70
Age	0.93*	1.03	0.91*	0.98	0.99	0.93	0.99
	(0.90–0.96)	(0.99–1.06)	(0.86–0.96)	(0.94–1.02)	(0.95–1.02)	(0.87–1.001)	(0.96–1.02)
	p < .0001	p = .052	p = .0006	p = .26	p = .48	p = .06	p = .36
Education	0.95	1.18	0.61	0.74	0.83	0.26*	1.37
	(0.53–1.69)	(0.74–1.89)	(0.31–1.20)	(0.40–1.37)	(0.44–1.59)	(0.09–0.78)	(0.87–2.13)
	p = .86	p = .49	p = .15	p = .34	p = .58	p = .02	p =.17
Neuroticism	1.10	1.67*	1.23	1.07	1.40	1.17	1.55*
	(0.85–1.41)	(1.28–2.20)	(0.88–1.72)	(0.72–1.59)	(0.98–1.99)	(0.63–2.18)	(1.25–1.92)
	p = .47	p = .0002	p = .23	p = .74	p = .06	p = .61	p < .0001
Novelty seeking	1.33	1.05	1.01	0.93	1.22	0.89	1.05
	(0.97–1.83)	(0.83–1.33)	(0.69–1.49)	(0.62–1.39)	(0.84–1.76)	(0.48–1.66)	(0.84–1.31)
	p = .07	p = .69	p = .96	p = .72	p = .30	p = .72	p = .65
Conduct disorder	1.00	0.36*	2.14*	2.37	1.71	0.75	0.61
	(0.60–1.66)	(0.15–0.88)	(1.16–3.96)	(0.87–6.46)	(0.86–3.39)	(0.28–2.04)	(0.27–1.37)
	p = .99	p = .025	p = .015	p = .09	p = .12	p = .58	p = .23
Primary depression	5.14*	12.40*	6.77*	1.35	2.38*	2.45	11.84*
	(1.92–13.73)	(4.96–31.04)	(1.33–34.50)	(0.60–3.06)	(1.24–4.55)	(0.88–6.76)	(6.00–23.35)
	p = .0011	p < .0001	p = .02	p = .46	p = .009	p = .09	p < .0001
Nicotine dependence	2.16*	2.01*	4.00*	2.04	1.64	1.74	1.75*
	(1.09–4.28)	(1.13–3.58)	(1.71–9.33)	(0.61–6.83)	(0.72–3.78)	(0.38–7.93)	(1.12–2.75)
	p = .03	p = .018	p = .001	p = .25	p = .24	p = .47	p=.01
Age at onset of alcohol use	0.88*	0.94	0.98	0.94	1.09	0.86	0.94
	(0.80–0.96)	(0.87–1.009)	(0.89–1.08)	(0.86–1.02)	(0.99–1.21)	(0.72–1.02)	(0.88–1.02)
	p = .006	p = .08	p = .63	p =.15	p = .09	p = .08	p=.13
Early drug use	4.11*	2.99*	3.61*	9.89*	6.53*	11.90*	5.77*
	(2.42–6.98)	(1.71–5.24)	(1.79–7.28)	(3.17–30.92)	(3.06–13.94)	(1.89–74.99)	(3.56–9.37)
	p < .0001	p = .0001	p = .0003	p < .0001	p < .0001	p = .008	p < .0001
Maternal AD	1.14	2.22*	1.27	0.65	1.80	3.26*	1.66*
	(0.61–2.13)	(1.29–3.80)	(0.60–2.67)	(0.22–1.93)	(0.86–3.79)	(1.25–8.48)	(1.02–2.69)
	p = .67	p = .004	p = .53	p = .44	p = .12	p = .02	p = .04
Paternal AD	1.04	1.34	1.91	0.78	1.18	0.88	1.13
	(0.59–1.80)	(0.82–2.20)	(1.00–3.67)	(0.35–1.75)	(0.62–2.26)	(0.34–2.27)	(0.73–1.76)
	p = .90	p = .25	p = .051	p = .55	p = .61	p = .79	p = .57
Primary Depression × Gendera	−1.19	−1.09	−1.06				−1.16
	p = .03	p = .04	p = .18				p = .009
Conduct Disorder × Gendera		1.96					1.23
		p = .0003					p = .01

Notes: Variables were coded as follows: conduct disorder, primary depression, nicotine dependence, maternal alcohol dependence (AD), paternal AD; 0 = no, 1 = yes. Gender: 0 = female, 1 = male. Neuroticism and novelty seeking: z scores. Age and age at onset of alcohol use: continuous. Early illicit drug use: 0 = age 18 or younger, 1 = after age 18 or never. Education: 0 = did not complete secondary school, 1 = completed secondary school. A blank space signifies that a variable was not included in the final model for a particular outcome because of its lack of significance in the univariate model for that outcome.

^aRegression weights and standard errors are displayed for interaction terms. Odds ratios for each stratum are reported in the results section.

^*Significant at α = .05.