Table 1.

Accuracy of Parameter Estimates for GGUM Concurrent Calibration With Three Anchor-Item Designs.

			Dichotomous								Polytomous
			α		δ		τ		θ		α		δ		τ		θ
Dist.	Design	N	RMSE	Bias	RMSE	Bias	RMSE	Bias	RMSE	Bias	RMSE	Bias	RMSE	Bias	RMSE	Bias	RMSE	Bias
Equivalent	Common-item	100	.59	.40	.84	.17	.63	−.37	.62	.17	.30	−.20	.47	.22	.48	−.01	.31	.16
		200	.42	.23	.56	.14	.61	−.40	.56	.04	.30	−.25	.44	.06	.48	−.10	.29	.04
		400	.30	.14	.51	.17	.40	−.24	.54	.07	.20	−.15	.28	.11	.31	−.11	.26	.08
		800	.21	.08	.49	.14	.34	−.13	.53	.07	.13	−.09	.18	.09	.27	−.09	.25	.07
	Standard	100	.59	.41	.52	.14	.64	−.39	.53	.17	.31	−.22	.48	.21	.48	.00	.31	.16
		200	.40	.22	.45	.07	.62	−.42	.46	.04	.30	−.26	.46	.06	.45	−.09	.29	.04
		400	.28	.14	.39	.07	.41	−.26	.45	.08	.21	−.17	.30	.11	.30	−.10	.26	.07
		800	.19	.07	.32	.06	.35	−.12	.45	.08	.14	−.10	.18	.11	.26	−.09	.25	.07
	Block-interlaced	100	.57	.39	.49	.14	.57	−.34	.52	.17	.29	−.19	.45	.19	.48	.01	.31	.16
		200	.37	.18	.41	.03	.55	−.37	.45	.04	.28	−.23	.43	.05	.44	−.09	.29	.04
		400	.27	.14	.39	.04	.38	−.22	.30	−.11	.18	−.13	.28	.07	.30	−.10	.26	.07
		800	.18	.07	.30	.05	.32	−.13	.45	.08	.13	−.09	.18	.07	.24	−.08	.24	.08
Non-equivalent	Common-item	100	.54	.31	.68	.16	.88	−.60	.47	.05	.36	−.31	.65	.07	.56	−.21	.31	.16
		200	.42	.22	.67	.16	.61	−.40	.64	−.12	.28	−.24	.46	−.10	.47	−.08	.31	−.11
		400	.28	.12	.54	.14	.58	−.28	.55	−.05	.25	−.21	.41	.04	.32	−.13	.34	−.04
		800	.18	.03	.41	.10	.41	−.21	.49	−.05	.18	−.14	.24	.02	.26	−.16	.27	−.05
	Standard	100	.54	.32	.59	.09	.87	−.63	.46	.05	.37	−.31	.66	.06	.56	−.20	.35	.05
		200	.40	.21	.44	−.02	.62	−.44	.47	−.11	.29	−.25	.48	−.09	.45	−.10	.31	−.11
		400	.27	.11	.38	.02	.57	−.34	.45	−.05	.25	−.22	.37	.01	.32	−.14	.29	−.05
		800	.17	.03	.31	−.02	.43	−.25	.45	−.05	.19	−.16	.27	.02	.25	−.15	.27	−.05
	Block-interlaced	100	.51	.28	.69	.15	.85	−.63	.45	.06	.36	−.31	.69	.06	.58	−.23	.35	.05
		200	.33	.15	.41	−.04	.55	−.38	.46	−.11	.27	−.22	.44	−.13	.42	−.09	.30	−.11
		400	.23	.07	.36	−.01	.46	−.29	.44	−.05	.21	−.18	.33	−.01	.38	−.15	.28	−.05
		800	.16	.01	.29	.01	.34	−.21	.45	−.05	.15	−.13	.24	.02	.23	−.17	.26	−.05

Note. GGUM = generalized graded unfolding model; Dist. = subpopulation trait distribution conditions; Design = anchor-item design; N = sample size per group; RMSE = root mean square error; Bias = bias of estimates; Equivalent = trait distribution for all groups is N(0, 1); Non-equivalent = trait distribution for Group 1 is N(0, 1) and trait distribution for Groups 2 and 3 is N(−0.5, 1).