. Author manuscript; available in PMC: 2023 Aug 1.

Published in final edited form as: J Subst Abuse Treat. 2022 Mar 22;139:108766. doi: 10.1016/j.jsat.2022.108766

Table 3.

Predictors of early and late therapeutic alliance in hierarchical stepwise linear regression models

							95% CI
	R ²	F	df	β	SE	p	LL	UL
Early alliance
Block 1	.198	3.126^*	3, 38
Age				−.242	.010	.161	−.036	.006
Education				.411	.003	.021	.001	.014
Income				.075	< .001	.660	< .001	< .001
Block 2	.257	1.428	8, 33
Baseline substance use (number of days)				.050	.043	.771	−.075	.100
Baseline PTSD symptom severity				−.061	.004	.727	−.010	.007
Early PTSD symptom improvement				.281	.568	.132	−.277	2.035
Intervention group (SS vs WHE)				.052	.164	.762	−.284	.384
Site				.055	< .001	.731	−.001	.001
Block 3	.394	2.315^*	9, 32
Number of alcohol/drug treatment attempts				.383	.006	.011	.004	.028
Late alliance^a
Block 1	.012	.089	3, 32
Age				.008	.025	.757	−.043	.059
Education				< .001	.008	.997	−.015	.015
Income				< .001	< .001	.651	< .001	.001
Block 2	.282	1.066	7, 28
Baseline substance use (number of days)				.077	.087	.375	−.093	.247
Baseline PTSD symptom severity				.004	.009	.674	−.014	.022
Early PTSD symptom improvement				.768	1.282	.551	−1.786	3.322
Intervention group (SS vs WHE)				.255	.312	.413	−.357	.868
Site				−.002	.001	.061	−.003	< .001

p < .05

Late alliance values for R², F, and degrees of freedom (df) are presented from original data. Values for individual variables presented are pooled unstandardized Betas and test statistics.

Note. SS = Seeking Safety; WHE = Women’s Health Education. Variables excluded for early and late alliance models: lifetime interpersonal problems, number of mental health treatment experiences, experiencing alcohol/drug or psychological/emotional problems, burden of alcohol/drug or psychological/emotional problems, and perceived treatment importance. Also excluded from late alliance model: number of alcohol/drug treatment attempts.