Comparison of Prior Setting Methods for Multilevel Model Effect Estimation Based on Small Sample Imbalanced Nested Data in Bayesian Framework

Table 3

Estimated accuracy results of treatment effects and nesting effects for the SSIND under all operating conditions (c = 16).

The number of subjects in each group	ICC	Prior setting	Convergence rate		Cover		Bias		RMSE
			Treatment	Nesting	Treatment	Nesting	Treatment	Nesting	Treatment	Nesting
			Effect	Effect	Effect	Effect	Effect	Effect	Effect	Effect
5	0.05	U	97.7	98.5	96.21	97.46	−45	105	80	30
		G	96.5	96.8	92.33	92.67	−6	−285	80	33
	0.10	U	97	98.5	95.88	96.04	−42	314	84	44
		G	96.5	97	92.12	94.23	−31	−171	84	29
	0.15	U	96.6	98.4	95.55	92.38	−43	563	88	64
		G	97.8	97	91.62	96.19	−23	−26	89	30
10	0.05	U	96.6	99	96.79	93.03	−23	−159	59	28
		G	96.9	97.4	94.22	73.1	−25	−361	60	38
	0.10	U	97.6	98.9	96.41	95.55	−35	102	65	32
		G	97.2	96.8	93.31	87.3	−25	−178	64	28
	0.15	U	97.8	98.7	95.91	91.79	−42	387	70	50
		G	97.5	97.1	92.92	92.48	−43	16	71	27
15	0.05	U	97.3	98.1	94.66	85.52	−37	−257	54	30
		G	97.9	97.2	92.65	64	−39	−391	54	40
	0.10	U	97.3	98.1	94.45	94.7	−54	28	61	28
		G	98.1	97.2	92.05	85.49	−40	−178	61	26
	0.15	U	97.7	98.3	95.5	92.47	−67	329	67	44
		G	97.5	97.7	92.82	93.14	−57	36	67	26
20	0.05	U	97.5	98.8	95.49	80.16	−25	−291	45	32
		G	97.4	98.3	94.56	57.58	−21	−398	46	40
	0.10	U	96.3	98.4	95.53	92.07	−30	5	52	26
		G	96.4	97.8	93.88	85.07	−33	−172	53	25
	0.15	U	98.1	98.6	95.72	92.9	−32	313	59	41
		G	97.4	97.7	93.53	93.14	−28	46	60	25