Table 4. Validity of thresholds other...

Table 4.

Validity of thresholds other than 0.95

	Estimation Model 2					Estimation Model 3
Cutoff	Estimated discontinuity (robust)			Effective number of observations		Estimated discontinuity (robust)			Effective number of observations
Cutoff	Coef.	Std. Err.	P > z	Left	Right	Coef.	Std. Err.	P > z	Left	Right
0.935	0.032	0.064	0.618	239	497	0.034	0.067	0.610	303	899
0.940	−0.142	0.066	0.032	426	694	−0.110	0.078	0.157	466	861
0.945	0.160	0.165	0.331	644	1,127	0.178	0.170	0.294	826	1,693
0.950	−0.454	0.132	0.001	1,139	1,486	−0.444	0.134	0.001	1,451	2,781
0.955	0.047	0.040	0.241	796	818	−0.062	0.056	0.268	1,015	1,076
0.960	0.002	0.012	0.899	723	1,014	0.005	0.013	0.724	813	1,158
0.965	−0.006	0.005	0.232	851	837	−0.003	0.004	0.461	1,046	1,092

	Estimation Model 2					Estimation Model 3
Cutoff	Estimated discontinuity (robust)			Effective number of observations		Estimated discontinuity (robust)			Effective number of observations
Cutoff	Coef.	Std. Err.	P > z	Left	Right	Coef.	Std. Err.	P > z	Left	Right
0.935	0.032	0.064	0.618	239	497	0.034	0.067	0.610	303	899
0.940	−0.142	0.066	0.032	426	694	−0.110	0.078	0.157	466	861
0.945	0.160	0.165	0.331	644	1,127	0.178	0.170	0.294	826	1,693
0.950	−0.454	0.132	0.001	1,139	1,486	−0.444	0.134	0.001	1,451	2,781
0.955	0.047	0.040	0.241	796	818	−0.062	0.056	0.268	1,015	1,076
0.960	0.002	0.012	0.899	723	1,014	0.005	0.013	0.724	813	1,158
0.965	−0.006	0.005	0.232	851	837	−0.003	0.004	0.461	1,046	1,092

Note(s):

This is a summary of the 14 estimated results for Models 2 and 3 in Table 3, with only the cutoff values changed. Only the “estimated discontinuity” and the “effective number of observations” are presented for each estimation model. The highlighted line at 0.950 is reproduced from Table 3 as a benchmark. The effective number of observations is the number of observations within the range of 0.95±bw

Source(s): Author’s own work

[ViewLarge]

Sharing Unavailable