Table 2.

Taxonomies of Nested Logistic Regression Models that Display the Fitted Relationship Between Youths’ Level of Depression Symptoms, Gender, and (a) the Proportion (All, Some, None) of Habitual Coping ResponsesMatching Evidence-Based Psychotherapy (EBP) Components (n = 104), (b) the Proportion of Habitual Coping Responses Matching Youths’ Identified Perceived-Effective Responses (n = 103), and (c) the Number (Three and At Least One) of Perceived-Effective Coping Responses Reported (n = 103)

	Models
Parameter	M0	MD	MG	MDG	MD×G
(a) Matching EBP Componentsa
Intercept 2		−0.05	−0.35	−0.41	−0.39
Intercept 1		1.82***	1.51***	1.51***	1.53***
Depression Symptoms		−0.40†		−0.46*	−0.28
Genderb			0.68†	0.78*	0.79*
Depression Symptoms × Gender					−0.35
−2LL	207.29	203.65	204.03	199.58	198.92
NagelkerkeR2	0.000	0.040	0.036	0.083	0.090
(b) Matching Youths’ Perceived-Effective Responsesa
Intercept 2		−1.53***	−1.81***	−1.88***	−1.88***
Intercept 1		0.23	−0.07	−0.07	−0.08
Depression Symptoms		−0.50*		−0.52*	−0.59†
Genderb			0.63†	0.68†	0.68†
Depression Symptoms × Gender					0.14
−2LL	214.17	208.81	211.32	205.62	205.52
NagelkerkeR2	0.000	0.058	0.031	0.091	0.092
(c) No. of Perceived-Effective Responsesc
Three Perceived-Effective Responses
Intercept		0.62**	0.69*	0.69*	0.71*
Depression Symptoms		−0.14		−0.14	−0.38
Genderb			−0.15	−0.14	−0.16
Depression Symptoms × Gender					0.49
−2LL	133.31	132.97	133.18	132.84	131.74
NagelkerkeR2	0.000	0.005	0.002	0.006	0.021
At Least One Perceived-Effective Response
Intercept		2.17***	1.75***	1.95***	1.92***
Depression Symptoms		−0.89**		−0.91**	−0.83†
Genderb			0.43	0.52	0.62
Depression Symptoms × Gender					−0.19
−2LL	78.10	70.73	77.60	70.06	69.99
NagelkerkeR2	0.000	0.130	0.009	0.141	0.143

Note. M₀ = null model with no predictors; M_D = model with depression symptoms as the only predictor; M_G = model with gender as the only predictor; M_DG = model with depression symptoms and gender as predictors; M_D×G = model with depression symptoms, gender, and their interaction as predictors. Significance levels were based on the Wald test. Regression coefficients for the predictors with p< .10 are in shown in boldface.

^aThe proportional odds assumption was met, thus three-level ordinal outcomes were used in analyses; the models predict youths reporting a higher proportion of habitual responses matching EBP components or perceived- effective responses.

^bGender was coded as 0 for girls and 1 for boys.

^cThe proportional odds assumption was not met, thus two binary outcomes were used.

^†p < .10;

^*p < .05;

^**p <.01,

^***p <.001.