Table 4.

Multivariate comparison of factors associated with students experiencing alcohol-related non-consensual sex, by sex

Variable	Category	Female students			Male students
		Adjusted odds ratio (95% confidence interval)	Wald	P	Adjusted odds ratio (95% confidence interval)	Wald	P
Alcohol consumption	Non hazardous	Reference	-	-	Not entered into the model
	Hazardous	2.82 (1.96-4.06)	30.85	< 0.001
A is mildly drunk, B severely. B is too drunk to give consent. They have sex. Next day B states rape occurred. Should A be held responsible for rape?	Disagree	Reference	7.82	0.020	Not entered into the model
	Undecided	0.93 (0.58-1.50)	0.08	0.774
	Agree	1.51 (1.03-2.19)	4.54	0.033
Being drunk when having sex increases the likelihood of a false rape allegation	Disagree	Reference	8.67	0.013	Not entered into the model
	Undecided	0.70 (0.36-1.36)	1.12	0.290
	Agree	0.53 (0.35-0.82)	8.34	0.004
Have sex with:	Opposite sex	Not entered into the model			Reference	10.26	0.006
	Same sex				2.80 (1.07-7.32)	4.39	0.036
	Both sexes				6.60 (1.60-27.17)	6.83	0.009
Consent is about having the freedom to choose to have sex	No	Not entered into the model			Reference	6.79	0.034
	Unsure				0.20 (0.05-0.76)	5.57	0.018
	Yes				0.84 (0.34-2.05)	0.15	0.696
A and B are severely drunk. A is too drunk to establish if consents present. B is too drunk to consent. Both have sex. Next day B states rape occurred. Should A be held responsible for rape? How would you describe this?	Consensual	Not entered into the model			Reference	5.69	0.128
	Rape				0	0	0.999
	A mid-point				0.99 (0.45-2.19)	0	0.986
	Undecided				0.08 (0.01-0.70)	5.26	0.022

Only variables (or variable strata) significantly related to the dependent variable are retained in the logistic regression model. N = 789 for the female logistic regression model. Hosmer and Lemeshow Goodness-of-Fit Test: X² = 1.18, P = 0.95. N = 250 for the male logistic regression model. Hosmer and Lemeshow Goodness-of-Fit Test: X² = 1.70, P = 0.95.