Table 3.

Final logistic regression models for factors associated with lifetime DSM-III-R depression symptoms, MDE, and MDD for boys (n=98) and girls (n=104).

Outcome Variable	Boys			Girls
	Factor*	Odds Ratio	p-value	Factor*	Odds Ratio	p-value
Depressed Mood	MJ Dep	3.506	0.026	Age	1.475	0.014
Anhedonia	MJ Dep	3.474	0.032	Age	1.440	0.016
Weight Change	MJ Dep	2.934	0.087	None
Sleep Change	Age	1.407	0.054
	MJ Dep	2.925	0.068	None
Psychomotor	MJ Dep	5.833	0.003	None
Loss of Energy	MJ Dep	2.934	0.087	Age	1.397	0.025
Worthless/Guilty				Early use	0.295	0.030
	MJ Dep	4.867	0.011	Heavy use	3.711	0.026
Poor Concentration	MJ Dep	3.474	0.032	None
Suicide				Early use	0.436	0.083
	Age	2.048	0.052	CD	3.73	0.037
Major Depression Episode (MDE)	Age	1.477	0.079
	MJ Dep	4.856	0.014	Age	1.395	0.027
Major Depression Disorder (MDD)	None			Age	1.33	0.056

Multivariate logistic regression was used to assess the association of each depression symptom, MDE, and MDD (as the outcome variable) with the following independent variables or factors entered into the initial model in each regression: age, gender, cannabis dependence, alcohol dependence, heavy substance use, early substance use, and early conduct disorder. Final logistic models were determined using the Backward Stepwise approach of Wald.

^*Factors in this column are those independent variables from the initial model which appeared in the final logistic regression model in each analysis. Alpha was set at 0.05, and p-values < 0.05 were considered significant.