. Author manuscript; available in PMC: 2012 Dec 19.

Published in final edited form as: Open Addict J. 2011 Nov 11;4:58–64. doi: 10.2174/1874941001104010058

Table 5.

Stepwise Multivariate Linear Regression of Predictors on MARS-8

	Unstandardized Beta	Standard Error	P-value
Step 1: Enter all Background Variables
Age	-.001	.002	.545
Male	-.001	.048	.984
Black	.010	.060	.872
Hispanic	-.018	.058	.759
Completed High School/GED	.016	.047	.731
Public Assistance/Disability	.068	.052	.191
Arrested as Adult	.096	.053	.071
Alcohol/Drug Use Intensity	.001	.001	.107
Drug Toxicology Positive R² = .08	-.066	.050	.193
Step 2: Add Significant Hypothesized Risk Factors^*
Age	-.002	.002	.304
Male	-.017	.048	.718
Black	.026	.057	.647
Hispanic	-.021	.055	.702
Completed High School	.010	.045	.820
Public Assistance/Disability	.073	.049	.141
Arrested as Adult	.044	.052	.398
Alcohol/Drug Use Intensity	.000	.001	.785
Drug Toxicology Positive	-.044	.049	.369
Medication Side Effects	.014	.005	.006
Self-Efficacy for Drug Avoidance	-.022	.016	.173
Satisfaction with Medication	.020	.026	.436
Recovery Support	-.032	.026	.224
Friend's Support for Abstinence	-.015	.036	.681
Recovery-Promoting Behaviors R² = .23	-.029	.029	.320
Step 3: Backward Elimination of Hypothesized Risk Factors
Age	-.002	.002	.286
Male	-.009	.047	.846
Black	.032	.057	.575
Hispanic	-.017	.055	.758
Completed High School	.005	.044	.910
Public Assistance/Disability	.076	.049	.124
Arrested as Adult	.050	.051	.324
Alcohol/Drug Use Intensity	.000	.001	.579
Drug Toxicology Positive	-.054	.048	.263
Medication Side Effects	.013	.005	.008
Self-Efficacy for Drug Avoidance	-.029	.014	.043
Recovery Support R² = .21	-.046	.023	.046

Risk factors for non-adherence that had significant bivariate correlations (shown in Table 4) were entered as a block.