Table 2.

The Type–I Error Simulation Results of the Chi-square and t statistics from 500 Replications. The Values are Percentage of the Detected Outliers.

Factors			Jointa	Independent		t statistic
S-to-N	T	n		Innb.	Addc.	η1	η2	y1	y2	y3	y4	y5	y6
High	60	100	.014	.010	.014	.008	.008	.008	.007	.008	.007	.007	.007
	60	300	.014	.010	.015	.008	.008	.008	.008	.007	.008	.007	.007
	100	60	.013	.010	.013	.009	.009	.008	.008	.009	.008	.008	.008
	300	60	.011	.010	.011	.010	.009	.009	.010	.009	.010	.010	.009
Low	60	100	.010	.010	.010	.008	.008	.007	.007	.008	.007	.007	.007
	60	300	.010	.010	.010	.008	.008	.007	.007	.007	.008	.007	.007
	100	60	.010	.010	.010	.009	.008	.009	.009	.008	.008	.009	.009
	300	60	.010	.010	.010	.010	.009	.010	.010	.009	.010	.009	.010

^a.Joint chi–square statistic (${\rho }_{it}^{*2}$)

^b.Independent chi–square statistic for innovative outlier (${\rho }_{\eta ,it}^{*2}$)

^c.Independent chi–square statistic for additive outlier (${\rho }_{y,it}^{*2}$)