Skip to main content
. Author manuscript; available in PMC: 2020 Mar 1.
Published in final edited form as: Alcohol Clin Exp Res. 2019 Feb 11;43(3):497–508. doi: 10.1111/acer.13960

Table 3.

Regression Models for Past-Year DSM-5 Alcohol Use Disorder Severity as a Function of Sexual Orientation Discrimination: Subpopulation of U.S. Sexual Minorities Asked About Sexual Orientation Discrimination

 Linear Regression  Negative Binomial Regression (log link)
Model 1: Past-Year DSM-5 AUD Severity as a function of Past-Year Sexual Orientation Discrimination; n = 3,463 Model 2: Past-Year DSM-5 AUD Severity as a function of PPY Sexual Orientation Discrimination; n = 3,460 Model 3: Past-Year DSM-5 AUD Severity as a function of Past-Year Sexual Orientation Discrimination; n = 3,463 Model 4: Past-Year DSM-5 AUD Severity as a function of PPY Sexual Orientation Discrimination; n = 3,460
Covariates Estimated Coefficient [95% CI] Estimated Coefficient [95% CI] Estimated Coefficient [95% CI] Estimated Coefficient [95% CI]
Intercept 0.02 [−0.24, 0.28] 0.01 [−0.25, 0.27] −1.99 [−2.53, −1.44]** −2.00 [−2.56, −1.44]**
Sexual orientation discrimination scale1 0.07 [0.02, 0.12]* 0.04 [<0.01, 0.07] 0.06 [0.02, 0.10]* 0.04 [0.01, 0.07]
Sex
 Male
 Female −0.12 [−0.30, 0.06] −0.12 [−0.30, 0.06] −0.19 [−0.35, −0.03] −0.19 [−0.35, −0.03]
Race/ethnicity
 White
 Black 0.18 [−0.06, 0.42] 0.19 [−0.05, 0.44] 0.12 [−0.08, 0.33] 0.14 [−0.07, 0.34]
 Hispanic −0.06 [−0.28, 0.16] −0.04 [−0.26, 0.18] −0.17 [−0.38, 0.04] −0.16 [−0.37, 0.05]
 Other −0.01 [−0.28, 0.27] <0.01 [−0.27, 0.27] −0.08 [−0.37, 0.21] −0.06 [−0.35, 0.22]
Age
 65+
 45–64 0.32 [0.15, 0.48]** 0.33 [0.16, 0.49]** 1.21 [0.76, 1.65]** 1.22 [0.78, 1.65]**
 25–44 0.73 [0.55, 0.91]** 0.74 [0.56, 0.92]** 1.65 [1.20, 2.09]** 1.66 [1.22, 2.10]**
 18–24 1.04 [0.83, 1.25]** 1.06 [0.85, 1.28]** 1.88 [1.44, 2.32]** 1.90 [1.47, 2.34]**
Education
 HS or Less
 GED/Some Coll. 0.18 [−0.01, 0.38] 0.18 [−0.02, 0.37] 0.19 [0.02, 0.36] 0.18 [0.01, 0.36]
 College Grad + 0.13 [−0.09, 0.34] 0.11 [−0.10, 0.33] 0.15 [−0.05, 0.36] 0.13 [−0.07, 0.34]
Urbanicity
 Urban
 Rural −0.21 [−0.45, 0.03] −0.21 [−0.45, 0.03] −0.29 [−0.60, 0.03] −0.28 [−0.60, 0.03]
Geographical region
 Northeast
 Midwest −0.05 [−0.32, 0.21] −0.04 [−0.31, 0.22] −0.02 [−0.29, 0.26] −0.01 [−0.29, 0.26]
 South −0.22 [−0.47, 0.03] −0.21 [−0.46, 0.04] −0.15 [−0.37, 0.07] −0.14 [−0.37, 0.09]
 West −0.01 [−0.26, 0.25] −0.01 [−0.26, 0.25] 0.08 [−0.16, 0.31] 0.07 [−0.17, 0.31]
Any other past-year substance use disorder (SUD)2
 No past-year SUD
 Any other past-year SUD 1.31 [1.08, 1.55]** 1.32 [1.09, 1.56]** 1.05 [0.88, 1.23]** 1.06 [0.88, 1.23]**
Lifetime conduct or antisocial disorder3
 No disorder
 Conduct/antisocial disorder 0.80 [0.31, 1.29]* 0.81 [0.32, 1.30]* 0.40 [0.16, 0.64]* 0.39 [0.15, 0.64]*
Household history of substance problem by parent/adult
 No household history
 Yes household history 0.11 [−0.12, 0.34] 0.10 [−0.13, 0.33] 0.15 [−0.03, 0.34] 0.16 [−0.03, 0.34]
 Don’t know/missing/other −0.20 [−1.16, 0.76] −0.20 [−1.16, 0.76] −0.57 [−1.58, 0.44] −0.58 [−1.59, 0.43]
Overdispersion Parameter (for Negative Binomial models)4 2.80 [2.49, 3.15]** 2.82 [2.50, 3.17]**
R-squared (for Linear Regression Models) 0.163 0.161

Notes. 95% CI = confidence interval.

1

The sexual orientation discrimination scale in model 1 and model 3 consisted of past-year sexual orientation discrimination experiences (0–24) while the sexual orientation discrimination scale in model 2 and model 4 consisted of prior-to-past-year sexual orientation discrimination experiences (0–24).

2

Any other past-year DSM-5 other substance use disorder consisted of cannabis, cocaine, heroin, hallucinogen, inhalant, prescription opioid, sedative/tranquilizer, stimulant, and/or other drug use disorder (e.g., ecstasy, ketamine).

3

DSM-5 conduct/antisocial personality disorder consisted of lifetime conduct disorder and/or antisocial personality disorder.

4

The overdispersion parameter captures the amount of additional variance above and beyond a Poisson distribution (where the mean of the DV is equal to the variance); if the reported confidence interval does not include zero, this suggests that the Negative Binomial model provides a better fit to the observed count data than the Poisson model.

= reference group.

*

p <= 0.01

**

p <= 0.001.