Table 4.

Multilevel Model Testing Associations Between Playing Drinking Games and the Number of Negative Alcohol Consequences.

	N1 = 670; N2 = 4,073
	OR/RR	95 % CI LL, UL
Negative Binomial Regression Submodel (Count)
Intercept	0.47**	0.32, 0.70
Daily-Level
Weekend	1.01	0.72, 1.41
Month	0.94	0.88, 1.00
Week	1.02	0.96, 1.08
Burst	1.05	0.98, 1.08
Number of drinks (CWP)	1.21***	1.18, 1.25
Played drinking game (CWP)	1.08	0.92, 1.27
Person-Level
Male Biological Sex	0.70**	0.56, 0.87
Age	0.83**	0.75, 0.93
Student	0.96	0.76, 1.22
Race
White, NH (reference)
Asian, NH	1.45*	1.06, 1.99
Black, NH	1.70**	1.15, 2.51
Hispanic	1.22	0.97, 1.54
Other	1.32	0.80, 2.18
Number of drinks (GMC)	1.41***	1.28, 1.54
Played drinking game (GMC)	0.95	0.88, 1.03
Logistic Regression Submodel (Likelihood)
Intercept	0.08	0, ∞
Daily-Level
Weekend	0.85	0.30, 2.40
Month	0.97	0.75, 1.25
Week	1.05	0.82, 1.35
Burst	1.27	0.99, 1.63
Number of drinks (CWP)	0.34***	0.25, 0.46
Played drinking game (CWP)	0	0, ∞
Person-Level
Male Biological Sex	0.84	0.44, 1.58
Age	1.14	0.87, 1.49
Student	1.09	0.58, 2.04
Race
White, NH (reference)
Asian, NH	0.37	0.13, 1.01
Black, NH	0.61	0.20, 1.92
Hispanic	0.81	0.43, 1.52
Other	0.70	0.24, 2.09
Number of drinks (GMC)	0.17***	0.09, 0.33
Played drinking game (GMC)	0.07	0, ∞

Note. Total number of negative consequences does not include the two driving-related outcomes. The estimation results of the model in the zero-inflation part was unstable, leading to very large standard errors; therefore, some of the upper limits in the confidence intervals approached infinity (∞). OR = odds ratio for logistic regression submodel; RR = rate ratio for negative binomial regression submodel; Daily-level predictors were centered-within-person (CWP); Person-level predictors were grand-mean-centered (GMC); LL = lower limit; UL = upper limit; NH = Non-Hispanic; N1 = Number of people; N2 = Number of days. * p <.05, ** p <.01, *** p <.001.