Skip to main content
. Author manuscript; available in PMC: 2019 Oct 1.
Published in final edited form as: Drug Alcohol Depend. 2018 Aug 18;191:215–222. doi: 10.1016/j.drugalcdep.2018.06.037

Table 4.

Alcohol Use Models

Model Effect b RR SE z p
Unmoderated Slope (relationship inv. → alcohol use)          
  Fixed Effect (average slope) −.44 .64 .02 −22.02 < .001
  Random Effect (variance of slope) .04   .01 3.09 .002
Gender Moderated Slope (relationship inv. → alcohol use)          
  Fixed Effect (average slope) −.32 .73 .01 −35.62 < .001
  Random Effect (variance of slope) .004   .002 2.17 .03
  Gender → Slope −.33   .01 −28.78 < .001
Sexual Identity Moderated Slope (relationship inv. → alcohol use)          
  Fixed Effect (average slope) −.12 .89 .02 −4.87 < .001
  Random Effect (variance of slope) .03   .01 6.43 < .001
  Sexual Identity → Slope −.01   .04 −.12 .91
Age Moderated Relationship Involvement −.07 .93 .09 −.76 .44
Age .05 1.05 .02 2.25 .02
Relationship Inv.*Age −.04 .96 .04 −.99 .32

Note. All coefficients were estimated controlling for age at the within-level and age, gender, sexual identity, and race/ethnicity at the between-persons level. Dichotomous variable coding: relationship involvement (not in a relationship = 0; in a relationship = 1), gender (cisgender women = 0; cisgender men = 1), sexual identity (lesbian/gay = 0; bisexual = 1). RR = rate ratio; relationship inv. = relationship involvement).