Table 3.

Wave 3 Items, Rates of Past 12-Month Endorsement, Item Response Model Parameters, and Differential Item Functioning by Assessment Year

	Wave 3 sample statistics					Longitudinal comparisona
			Model Fit			Wave 3 (18–24 years)	Wave 1 (12—18 years)
Item content (least to most severe)	% (n = 4,761)	EST (nm = 3,094)	SE	Infit	PTBS	(nm = 3,093) EST	(nm = 2,298) EST	DIF
12 or more drinks	58.9	–4.18	0.07	1.05	0.26	–3.97	–2.69	–1.28*
Gotten drunk	50.6	–2.83	0.05	0.88	0.45	–2.63	–2.44	–0.19
5 drinks in a row	48.6	–2.57	0.05	0.98	0.43	–2.37	–2.17	–0.20
Hung over	41.6	–1.82	0.04	0.94	0.48	–1.62	–0.95	–0.67*
Threw up	34.1	–1.07	0.04	1.12	0.41	–0.88	–0.85	–0.03
Monthly	44.6	–2.10	0.05	0.97	0.47	–1.90	–0.81	–1.09*
Weekly	28.6	–0.52	0.04	0.92	0.50	–0.33	0.47	–0.80*
Date problem	10.1	1.54	0.06	0.97	0.35	1.74	0.65	1.09*
Regretted sexual situation	14.5	0.94	0.05	1.03	0.38	1.13	0.76	0.37
Physical fight	7.3	2.03	0.06	1.04	0.31	2.23	1.16	1.07*
Friend problem	8.1	1.88	0.06	0.88	0.38	2.08	1.18	0.90*
School problem	5.5	2.44	0.07	0.91	0.34	2.64	2.12	0.52*
Drunk at school or work	4.9	2.59	0.07	1.01	0.28	–	–	–
Daily	2.0	3.67	0.11	1.12	0.15	3.88	3.57	0.31

EST, Rasch model severity estimate; DIF, differential item functioning; TBS, point biserial correlations; n_m, the number of participants with at least one but not all symptoms, who thereby are included in the Rasch model.

Infit values of 1 indicate perfect fit to the Rasch model; lower values indicate less random variation than expected and higher values indicate more random variation.

^aLongitudinal comparison estimates vary slightly from the estimates from the cross-sectional analyses because analyses were conducted using only those items that were common to both waves. Prevalence for each item remains unchanged and is not presented.

^*DIF that was statistically significant at the 0.05 level and above a moderate effect size cut-off point of 0.50.