Table 6 Necessary-condition effect sizes...

Table 6

Necessary-condition effect sizes and their sensitivity to the removal of influential observations

Condition (driver)	No. of outliers	d (original)	p-value	d (without outliers)	p-value	Change
KNOW	1	0.044	<0.001	0.067	<0.001	50.00%
OPTI	3	0.149	0.028	0.314	<0.001	110.60%
INNOV	0	<0.001	>0.999	<0.001	>0.999	–
ECU	2	0.075	<0.001	0.083	<0.001	10.20%

Condition (driver)	No. of outliers	d (original)	p-value	d (without outliers)	p-value	Change
KNOW	1	0.044	<0.001	0.067	<0.001	50.00%
OPTI	3	0.149	0.028	0.314	<0.001	110.60%
INNOV	0	<0.001	>0.999	<0.001	>0.999	–
ECU	2	0.075	<0.001	0.083	<0.001	10.20%

Condition (barrier)	No. of outliers	d (original)	p-value	d (without outliers)	p-value	Change
SCEP	3	0.045	0.014	0.088	<0.001	95.70%
INSEC	0	<0.001	>0.999	<0.001	>0.999	–
RISK	0	<0.001	>0.999	<0.001	>0.999	–
NREG	0	<0.001	>0.999	<0.001	>0.999	–

Condition (barrier)	No. of outliers	d (original)	p-value	d (without outliers)	p-value	Change
SCEP	3	0.045	0.014	0.088	<0.001	95.70%
INSEC	0	<0.001	>0.999	<0.001	>0.999	–
RISK	0	<0.001	>0.999	<0.001	>0.999	–
NREG	0	<0.001	>0.999	<0.001	>0.999	–

Note(s): d denotes the necessary-condition effect size estimated using the ceiling envelopment–free disposal hull (CE-FDH) method. Statistical significance was assessed using permutation tests with 10,000 replications. Theoretical drivers were analysed using the upper-left ceiling, whereas theoretical barriers were analysed using the upper-right ceiling

Source(s): Author’s own work

[ViewLarge]

Sharing Unavailable