Table 5.

Results for average autoregressive parameter ϕ and the effect of a level 2 predictor z_i on the autoregressive parameter ϕ_i.

N	T	Bias				CR0.95
		γ10		γ11		γ10		γ11
		NC	CMC	NC	CMC	NC	CMC	NC	CMC
20	20	−0.007	−0.076	−0.050	−0.007	0.897	0.720	0.944	0.958
	50	−0.005	−0.030	−0.045	−0.026	0.922	0.856	0.939	0.944
	100	0.000	−0.013	−0.019	−0.008	0.923	0.909	0.942	0.944
50	20	0.004	−0.071	−0.068	−0.022	0.885	0.476	0.950	0.959
	50	0.001	−0.026	−0.032	−0.014	0.904	0.781	0.945	0.948
	100	0.000	−0.013	−0.018	−0.006	0.924	0.890	0.953	0.949
100	20	0.004	−0.071	−0.084	−0.036	0.918	0.170	0.921	0.940
	50	0.000	−0.027	−0.024	−0.003	0.907	0.628	0.944	0.955
	100	0.001	−0.012	−0.019	−0.008	0.928	0.832	0.939	0.942

Bias and coverage rate of 95% confidence intervals (CR_0.95) based on 1000 replications. Results for γ₁₀ = 0.3, that is, the average autoregressive parameter ϕ, and for γ₁₁ = 0.4, that is, the effect of a level 2 predictor on the autoregressive parameter, using NC and CMC (with sample mean) for the autoregressive predictor.