. Author manuscript; available in PMC: 2018 Nov 1.

Published in final edited form as: Psychol Addict Behav. 2017 Sep 21;31(7):775–785. doi: 10.1037/adb0000314

Table 5.

Multivariable Multinomial Logistic Regression Associations between Total Number of Alcohol Use Reason/Context Endorsements and Drinking Category among U.S. 12^th Grade Alcohol Users, 2005–2016

	Model 1^a			Model 2^b

	<5 Drinks [vs. 5–9]	10–14 Drinks [vs. 5–9]	15+ Drinks [vs. 5–9]	10–14 Drinks [vs. 15+]
	ARR (95% CI) p	ARR (95% CI) p	ARR (95% CI) p	ARR (95% CI) p

Total number of reasons	0.89 (0.87, 0.91) ^***	0.99 (0.96, 1.03)	1.04 (1.00, 1.08) ^*	0.95 (0.91, 0.99) ^*
Total number of contexts	0.67 (0.63, 0.69) ^***	1.31 (1.21, 1.41) ^***	1.81 (1.67, 1.97) ^***	0.72 (0.66, 0.78) ^***

Notes: N (unweighted) = 16,902. ARR = adjusted relative risk ratio; CI = confidence interval. Models controlled for total reasons and contexts simultaneously, as well as covariates listed in Table 1 (estimates for covariates not shown).

p<0.05;

^**

p<0.01;

^***

p<0.001.

Model 1 results use 5–9 drinks as the base class.

Model 2 results use 15+ drinks as the base class to enable comparison of the likelihoods of 10–14 versus 15+ drinks by reason/context endorsement.